1 00:00:04,480 --> 00:00:12,400 Speaker 1: Welcome to tech Stuff, a production from iHeartRadio. Hey there, 2 00:00:12,440 --> 00:00:16,320 Speaker 1: and welcome to tech Stuff. I'm your host, Jonathan Strickland. 3 00:00:16,480 --> 00:00:20,000 Speaker 1: I'm an executive producer with iHeart Podcasts and how the 4 00:00:20,040 --> 00:00:23,560 Speaker 1: tech are you? So what you're about to hear is 5 00:00:23,800 --> 00:00:27,880 Speaker 1: an experiment. I decided to see what chat GPT could 6 00:00:27,960 --> 00:00:31,200 Speaker 1: scramble up if I asked it to create a tech 7 00:00:31,440 --> 00:00:35,720 Speaker 1: podcast episode. Now, I didn't specifically say a tech stuff 8 00:00:35,880 --> 00:00:41,080 Speaker 1: episode in particular. Instead, my prompt literally read quote, create 9 00:00:41,240 --> 00:00:46,440 Speaker 1: a tech podcast episode about airbags, end quote, and I 10 00:00:46,560 --> 00:00:50,840 Speaker 1: just chose airbags at random. This is what it came 11 00:00:50,960 --> 00:00:54,639 Speaker 1: up with. So, without further ado, listeners, I bring you 12 00:00:55,560 --> 00:00:59,760 Speaker 1: the following episode from chat GPT, and this will all 13 00:00:59,840 --> 00:01:03,880 Speaker 1: be quoted directly from chat GPT. You'll know when I 14 00:01:03,960 --> 00:01:08,840 Speaker 1: come back in. So here we go. Tech Innovations, air 15 00:01:08,920 --> 00:01:13,720 Speaker 1: Bags and Automotive Safety. Host Welcome to Tech Innovations, the 16 00:01:13,760 --> 00:01:16,720 Speaker 1: podcast where we explore the latest and greatest in technology 17 00:01:16,720 --> 00:01:21,480 Speaker 1: and innovation. I'm your host, your name, and today we're 18 00:01:21,520 --> 00:01:25,920 Speaker 1: diving into a crucial aspect of automotive safety. Airbags. These 19 00:01:26,000 --> 00:01:29,039 Speaker 1: life saving devices have been a staple of vehicles for decades, 20 00:01:29,120 --> 00:01:32,800 Speaker 1: but their technology and impact are often overlooked. Today, we'll 21 00:01:32,840 --> 00:01:37,480 Speaker 1: uncover the history, technology, and future of airbags. Let's get started. 22 00:01:38,040 --> 00:01:42,399 Speaker 1: Segment one, The history of airbags. HOST. Air bags have 23 00:01:42,480 --> 00:01:45,360 Speaker 1: a fascinating history that dates back to the nineteen fifties. 24 00:01:45,480 --> 00:01:48,040 Speaker 1: The idea of a cushion to protect drivers in a 25 00:01:48,080 --> 00:01:51,800 Speaker 1: crash was first conceived by John W. Hetrick, who filed 26 00:01:51,840 --> 00:01:55,040 Speaker 1: the initial patent in nineteen fifty one. However, it wasn't 27 00:01:55,080 --> 00:01:57,640 Speaker 1: until the nineteen seventies the airbags started to become a 28 00:01:57,680 --> 00:02:01,680 Speaker 1: practical reality. Guest expert joining us today is doctor Emily 29 00:02:01,800 --> 00:02:05,680 Speaker 1: Roberts and automotive safety historian Doctor Roberts. Can you tell 30 00:02:05,760 --> 00:02:08,080 Speaker 1: us about the early development of air bags in their 31 00:02:08,080 --> 00:02:13,360 Speaker 1: initial reception? Dtor Roberts. Absolutely, The early prototypes were quite 32 00:02:13,440 --> 00:02:17,280 Speaker 1: rudimentary compared to today's standards. They relied on compressed air 33 00:02:17,440 --> 00:02:21,960 Speaker 1: and were often bulky and unreliable. However, as the technology evolved, 34 00:02:22,120 --> 00:02:24,880 Speaker 1: especially with the advent of solid state crash sensors in 35 00:02:24,919 --> 00:02:28,240 Speaker 1: the nineteen eighties, airbags became more efficient and began to 36 00:02:28,440 --> 00:02:35,720 Speaker 1: gain acceptance among manufacturers and consumers. Segment two. How airbags work. HOST. 37 00:02:36,320 --> 00:02:39,760 Speaker 1: Modern airbags are marvels of engineering at their core. They're 38 00:02:39,800 --> 00:02:43,080 Speaker 1: designed to deploy within milliseconds of a collision to cushion 39 00:02:43,120 --> 00:02:46,880 Speaker 1: and protect the occupants. But what exactly happens during those 40 00:02:46,960 --> 00:02:51,440 Speaker 1: critical moments? Guest expert to explain the technical workings, we 41 00:02:51,520 --> 00:02:56,280 Speaker 1: have James Lee, and automotive engineer specializing in safety systems. James, 42 00:02:56,480 --> 00:02:59,480 Speaker 1: can you break down the process for us? James Lee? 43 00:03:00,160 --> 00:03:03,160 Speaker 1: Or When a crash sensor detects a collision, it sends 44 00:03:03,160 --> 00:03:06,400 Speaker 1: an electric signal to the airbag control unit. This unit 45 00:03:06,400 --> 00:03:09,600 Speaker 1: triggers an igniter that rapidly inflates the air bag using 46 00:03:09,639 --> 00:03:14,280 Speaker 1: a chemical reaction, typically sodium azide. The airbag then deploys 47 00:03:14,360 --> 00:03:16,320 Speaker 1: at speeds of up to two hundred miles per hour, 48 00:03:16,440 --> 00:03:19,720 Speaker 1: providing a cushion to reduce the impact on the occupants. 49 00:03:20,160 --> 00:03:25,239 Speaker 1: Segment three types of airbags host airbags have evolved beyond 50 00:03:25,400 --> 00:03:29,200 Speaker 1: just standard front airbags. There are now side airbags, curtain 51 00:03:29,280 --> 00:03:32,600 Speaker 1: air bags, knee air bags, and more. Each type serves 52 00:03:32,600 --> 00:03:35,400 Speaker 1: a specific purpose in protecting different parts of the body 53 00:03:35,520 --> 00:03:39,160 Speaker 1: during various kinds of collisions. Guest Expert, we're joined by 54 00:03:39,200 --> 00:03:42,960 Speaker 1: Lisa Fernandez and expert and vehicle safety systems. Lisa, could 55 00:03:43,040 --> 00:03:45,600 Speaker 1: you tell us about the different types of air bags 56 00:03:45,600 --> 00:03:49,640 Speaker 1: in their roles Lisa Fernandez. Of course, front air bags 57 00:03:49,640 --> 00:03:51,720 Speaker 1: are designed to protect the head and chest in a 58 00:03:51,760 --> 00:03:55,480 Speaker 1: frontal crash. Side airbags help protect the torso and pelvis 59 00:03:55,520 --> 00:03:58,640 Speaker 1: and side impacts, while curtain air bags deploy from the 60 00:03:58,720 --> 00:04:01,800 Speaker 1: roof to shield the head in rollovers. Knee airbags are 61 00:04:01,800 --> 00:04:05,760 Speaker 1: intended to prevent leg injuries. Together they create a comprehensive 62 00:04:05,800 --> 00:04:09,920 Speaker 1: safety net. Segment four. The impact of airbags on safety 63 00:04:09,960 --> 00:04:13,960 Speaker 1: statistics HOST. The introduction and widespread adoption of airbags have 64 00:04:14,080 --> 00:04:17,520 Speaker 1: had a profound impact on vehicle safety. According to the 65 00:04:17,600 --> 00:04:22,880 Speaker 1: National Highway Traffic Safety Administration or NHTSA, airbags have saved 66 00:04:23,080 --> 00:04:27,000 Speaker 1: tens of thousands of lives since their introduction. Guest expert, 67 00:04:27,200 --> 00:04:31,960 Speaker 1: we have doctor Michael Harris, a statistician with the NHTSA, 68 00:04:32,040 --> 00:04:35,440 Speaker 1: to provide insights into the data. Doctor Harris, what do 69 00:04:35,520 --> 00:04:39,400 Speaker 1: the numbers tell us about the effectiveness of airbags, Doctor Harris? 70 00:04:39,560 --> 00:04:43,279 Speaker 1: The statistics are compelling. Front airbags have been credited with 71 00:04:43,400 --> 00:04:47,200 Speaker 1: reducing driver fatalities by about twenty nine percent and front 72 00:04:47,279 --> 00:04:51,839 Speaker 1: passenger fatalities by thirty two percent in frontal crashes. Side 73 00:04:51,839 --> 00:04:55,440 Speaker 1: airbags with head protection have shown even greater potential, reducing 74 00:04:55,480 --> 00:05:01,040 Speaker 1: deaths in side impacts by about thirty seven percent. Future 75 00:05:01,160 --> 00:05:04,960 Speaker 1: of airbag technology. Host. As we look to the future, 76 00:05:05,040 --> 00:05:09,480 Speaker 1: airbag technology continues to advance. Innovations such as pedestrian air 77 00:05:09,520 --> 00:05:12,640 Speaker 1: bags and smart air bags that adjust deployment based on 78 00:05:12,680 --> 00:05:15,880 Speaker 1: the occupant's size and position are now on the horizon. 79 00:05:16,040 --> 00:05:19,400 Speaker 1: Guest expert. To discuss these future trends, we have doctor 80 00:05:19,480 --> 00:05:23,839 Speaker 1: Sarah Kim, a researcher in automotive technology. Doctor Kim, what 81 00:05:24,040 --> 00:05:28,560 Speaker 1: exciting developments can we expect in air bag technology? Doctor Kim. 82 00:05:28,880 --> 00:05:31,359 Speaker 1: One of the most promising areas is the integration of 83 00:05:31,400 --> 00:05:35,520 Speaker 1: advanced sensors and artificial intelligence. These systems can assess the 84 00:05:35,520 --> 00:05:39,480 Speaker 1: severity of a crash and adjust airbag deployment accordingly. Additionally, 85 00:05:39,640 --> 00:05:42,960 Speaker 1: external air bags that deploy to protect pedestrians and cyclists 86 00:05:43,040 --> 00:05:47,040 Speaker 1: are being tested, which could significantly enhance overall road safety. 87 00:05:47,400 --> 00:05:52,479 Speaker 1: Closing remarks, Host, it's clear that airbags have revolutionized automotive 88 00:05:52,480 --> 00:05:56,000 Speaker 1: safety and continue to evolve from their humble beginnings to 89 00:05:56,040 --> 00:05:59,080 Speaker 1: the sophisticated systems we have today. Air Bags play a 90 00:05:59,240 --> 00:06:02,560 Speaker 1: crucial role in protecting lives. Thank you to our experts 91 00:06:02,560 --> 00:06:05,320 Speaker 1: for sharing their insights, and thank you to our listeners 92 00:06:05,320 --> 00:06:09,960 Speaker 1: for tuning in, Stay safe and stay curious. Host. This 93 00:06:10,040 --> 00:06:13,719 Speaker 1: concludes today's episode of tech Innovations. If you enjoyed this episode, 94 00:06:13,720 --> 00:06:16,120 Speaker 1: please subscribe and leave a review. If you have any 95 00:06:16,200 --> 00:06:19,120 Speaker 1: questions or topics you'd like us to cover in future episodes, 96 00:06:19,160 --> 00:06:21,960 Speaker 1: feel free to reach us on social media or via 97 00:06:22,080 --> 00:06:26,560 Speaker 1: our website. Until next time, I'm your name, and this 98 00:06:26,640 --> 00:06:30,520 Speaker 1: is tech Innovations. Okay, there we go. That's it. That 99 00:06:30,760 --> 00:06:36,520 Speaker 1: is the chat GPT generated episode about airbags. Obviously very 100 00:06:36,600 --> 00:06:39,799 Speaker 1: light episode, right, not like a tech stuff episode, doesn't 101 00:06:39,839 --> 00:06:42,440 Speaker 1: last forty seven minutes. In fact, it lasted less than 102 00:06:42,480 --> 00:06:46,960 Speaker 1: seven minutes with me reading through that, although I imagine 103 00:06:47,000 --> 00:06:49,560 Speaker 1: that the different segments would probably have things like musical 104 00:06:49,640 --> 00:06:52,880 Speaker 1: transitions and such if this were a fully produced episode, 105 00:06:52,880 --> 00:06:55,719 Speaker 1: and probably each of those segments was meant to be 106 00:06:55,839 --> 00:06:58,240 Speaker 1: fleshed out right like that was not meant to be 107 00:06:58,240 --> 00:07:01,599 Speaker 1: a full episode. It was more like the outline of 108 00:07:01,640 --> 00:07:06,719 Speaker 1: an episode. And after I read this, I wanted to 109 00:07:06,760 --> 00:07:10,080 Speaker 1: know what sources the AI actually used in order to 110 00:07:10,120 --> 00:07:12,800 Speaker 1: craft this episode. So this is what it had to 111 00:07:12,840 --> 00:07:16,000 Speaker 1: say when I asked for that information. The episode script 112 00:07:16,080 --> 00:07:21,240 Speaker 1: was created based on general knowledge about airbags and their history, technology, 113 00:07:21,280 --> 00:07:24,280 Speaker 1: and impact on automotive safety. The information is drawn from 114 00:07:24,400 --> 00:07:29,080 Speaker 1: widely known facts and common sources in the automotive safety field, 115 00:07:29,400 --> 00:07:34,440 Speaker 1: such as National Highway Traffic Safety Administration in HTSA reports 116 00:07:34,480 --> 00:07:38,520 Speaker 1: and statistics on airbag effectiveness and safety data, automotive safety 117 00:07:38,560 --> 00:07:42,920 Speaker 1: engineering principles, including how airbags function and their deployment mechanisms, 118 00:07:43,400 --> 00:07:46,920 Speaker 1: historical records and patents such as the original airbag patent 119 00:07:47,000 --> 00:07:49,840 Speaker 1: by John W. Hetrick. More on that in a minute, 120 00:07:50,040 --> 00:07:53,760 Speaker 1: Current trends in automotive technology, including advancements in smart air 121 00:07:53,800 --> 00:07:58,120 Speaker 1: bags and pedestrian protection systems. For more detailed or specific information, 122 00:07:58,200 --> 00:08:02,760 Speaker 1: here are a few potential sources that could provide the 123 00:08:02,880 --> 00:08:07,200 Speaker 1: data used in the episode. NHTSA website and their publications 124 00:08:07,200 --> 00:08:11,560 Speaker 1: on vehicle safety features, automotive engineering textbooks and technical papers 125 00:08:11,600 --> 00:08:15,600 Speaker 1: on airbag technology, interviews and articles from industry experts and 126 00:08:15,680 --> 00:08:19,280 Speaker 1: researchers in automotive safety, automotive history books that cover the 127 00:08:19,320 --> 00:08:22,440 Speaker 1: development and evolution of vehicle safety features. If you need 128 00:08:22,520 --> 00:08:26,040 Speaker 1: specific references or more detailed information from these sources, I 129 00:08:26,080 --> 00:08:29,280 Speaker 1: can look them up for you. Okay, I didn't bother 130 00:08:29,680 --> 00:08:32,320 Speaker 1: to go into deeper detail here. Maybe I should have, 131 00:08:32,840 --> 00:08:35,480 Speaker 1: but already this is a little puzzling, right because it 132 00:08:35,520 --> 00:08:39,040 Speaker 1: says here are a few potential sources that could provide 133 00:08:39,080 --> 00:08:42,599 Speaker 1: the data used in this episode, which is pretty darn ambiguous. 134 00:08:42,640 --> 00:08:45,880 Speaker 1: You know, the sources could do that, but there's no 135 00:08:45,960 --> 00:08:49,600 Speaker 1: guarantee they will, or that they actually served as a 136 00:08:49,679 --> 00:08:53,080 Speaker 1: crucial source for the data that was ultimately pulled for 137 00:08:53,120 --> 00:08:58,320 Speaker 1: this AI generated episode. The plot thickens now. I would say, 138 00:08:59,040 --> 00:09:05,280 Speaker 1: on the surface, this rudimentary tech podcast episode is more 139 00:09:05,400 --> 00:09:09,120 Speaker 1: or less fine, but let's dive down a bit and 140 00:09:09,200 --> 00:09:13,160 Speaker 1: double check the homework assignment, because there are problems here, 141 00:09:13,240 --> 00:09:16,640 Speaker 1: some of which are factual in nature and a lot 142 00:09:16,679 --> 00:09:19,559 Speaker 1: of which I would argue are more ethical. So the 143 00:09:19,600 --> 00:09:24,559 Speaker 1: AI generated episode is correct that John W. Hettrick frequently 144 00:09:24,600 --> 00:09:28,720 Speaker 1: gets the credit for inventing the airbag. The Institute of Physics, 145 00:09:28,720 --> 00:09:32,240 Speaker 1: which in turn cites sources such as D. Sherman's article 146 00:09:32,480 --> 00:09:36,280 Speaker 1: The Rough Road to Airbags, gives a slightly different date. However, 147 00:09:36,600 --> 00:09:39,840 Speaker 1: the AI had said that Hettrick filed for this patent 148 00:09:39,880 --> 00:09:43,520 Speaker 1: back in nineteen fifty one, whereas the Institute of Physics 149 00:09:43,840 --> 00:09:47,280 Speaker 1: says he actually got the idea in nineteen fifty two, 150 00:09:47,400 --> 00:09:50,240 Speaker 1: after he got into an automobile accident with his wife 151 00:09:50,240 --> 00:09:53,120 Speaker 1: and child in the car. So why would he patent 152 00:09:53,240 --> 00:09:55,719 Speaker 1: something in nineteen fifty one if he didn't even get 153 00:09:55,720 --> 00:09:58,559 Speaker 1: the idea for it until nineteen fifty two. Well, this 154 00:09:58,640 --> 00:10:01,160 Speaker 1: one is easy to solve. All we have to do 155 00:10:01,320 --> 00:10:05,200 Speaker 1: is pop over to a patent database and search for 156 00:10:05,360 --> 00:10:08,400 Speaker 1: Hendrick's name in the inventor field and see what it 157 00:10:08,480 --> 00:10:11,600 Speaker 1: has on there. When I did that, I saw that 158 00:10:11,720 --> 00:10:15,200 Speaker 1: the date for the initial filing of a patent from 159 00:10:15,320 --> 00:10:21,800 Speaker 1: John Hendrick about airbags was on August fifth, nineteen fifty two. Okay, 160 00:10:22,160 --> 00:10:25,080 Speaker 1: that's one data point that's wrong, but it's only wrong 161 00:10:25,160 --> 00:10:29,080 Speaker 1: by a year, right, Like, this isn't the worst mistake 162 00:10:29,320 --> 00:10:32,640 Speaker 1: that AI could make. It's the right person, it's the 163 00:10:32,720 --> 00:10:36,839 Speaker 1: right invention, it's just the wrong year. And goodness knows, 164 00:10:36,880 --> 00:10:38,880 Speaker 1: there are a lot of tech topics that I cover 165 00:10:39,040 --> 00:10:41,960 Speaker 1: in the show where I can't find a reliable source 166 00:10:42,000 --> 00:10:45,800 Speaker 1: to give me a definitive date for certain things. The 167 00:10:45,840 --> 00:10:48,040 Speaker 1: further back you go, the harder it is to do 168 00:10:48,080 --> 00:10:52,280 Speaker 1: that unless someone has specifically filed a document. But in 169 00:10:52,280 --> 00:10:55,079 Speaker 1: this case, we do have that. We have an official 170 00:10:55,240 --> 00:10:58,760 Speaker 1: filed document. In fact, Chad GBT said that the patent 171 00:10:58,920 --> 00:11:02,160 Speaker 1: could be one of the sources that pulled from it. 172 00:11:02,200 --> 00:11:06,000 Speaker 1: Obviously wasn't because if they had pulled from that source, 173 00:11:06,160 --> 00:11:09,000 Speaker 1: it would have said nineteen fifty two, not nineteen fifty one. 174 00:11:09,200 --> 00:11:12,800 Speaker 1: This was a claim that was easily verifiable. The AI 175 00:11:12,960 --> 00:11:16,360 Speaker 1: did not do that, But no big deal. Human writers 176 00:11:16,480 --> 00:11:18,880 Speaker 1: make these kind of mistakes all the time, right, like 177 00:11:19,200 --> 00:11:22,360 Speaker 1: they just don't take the extra effort. Heck, there are 178 00:11:22,520 --> 00:11:24,720 Speaker 1: episodes of tech stuff that have gone out there where 179 00:11:24,720 --> 00:11:27,480 Speaker 1: I have made similar errors because I did not take 180 00:11:27,520 --> 00:11:30,760 Speaker 1: the extra step to actually verify something that's on me. 181 00:11:31,120 --> 00:11:34,080 Speaker 1: So I don't want to hold AI to too high 182 00:11:34,080 --> 00:11:37,600 Speaker 1: a standard, but it's already proven to not be one 183 00:11:37,679 --> 00:11:41,360 Speaker 1: hundred percent reliable, which is an issue if it's being 184 00:11:41,440 --> 00:11:45,520 Speaker 1: presented as being one hundred percent reliable. Right not to 185 00:11:45,559 --> 00:11:47,840 Speaker 1: say that AI has to be better than humans, but 186 00:11:47,880 --> 00:11:50,199 Speaker 1: at least has to be as good as it's claim 187 00:11:50,280 --> 00:11:53,680 Speaker 1: to be. Well, what happened next is I would argue 188 00:11:54,200 --> 00:11:58,160 Speaker 1: a very big deal and a real issue, more so 189 00:11:58,320 --> 00:12:02,160 Speaker 1: than the fudging a year. That is because the AI 190 00:12:02,320 --> 00:12:05,000 Speaker 1: creates an expert and it wouldn't be the last time 191 00:12:05,400 --> 00:12:07,880 Speaker 1: the spoiler alert, and I'll say it again when we 192 00:12:07,920 --> 00:12:11,480 Speaker 1: get to each one, but all of the guest experts 193 00:12:11,760 --> 00:12:15,640 Speaker 1: appear to me at least to be invented. I do 194 00:12:15,679 --> 00:12:20,080 Speaker 1: not believe doctor Emily Roberts exists. Now. There are plenty 195 00:12:20,080 --> 00:12:23,720 Speaker 1: of people named Emily Roberts. There's even one who's named 196 00:12:23,760 --> 00:12:27,920 Speaker 1: Emily Roberts who's a child Passenger Safety technician or CPST, 197 00:12:28,440 --> 00:12:31,720 Speaker 1: and she creates content for TikTok and YouTube shorts about 198 00:12:31,760 --> 00:12:35,040 Speaker 1: the proper way to ensure children are protected while riding 199 00:12:35,040 --> 00:12:39,240 Speaker 1: in vehicles. So she's at least in the ballpark for 200 00:12:39,760 --> 00:12:44,680 Speaker 1: an expert who's talking about automotive safety features. But there is, 201 00:12:44,760 --> 00:12:48,280 Speaker 1: as far as I can determine, no automotive safety historian 202 00:12:48,840 --> 00:12:53,080 Speaker 1: named doctor Emily Roberts, at least not anyone I could find. However, 203 00:12:53,120 --> 00:12:57,120 Speaker 1: that's probably for the best, right because unless chat GPT 204 00:12:57,559 --> 00:13:02,559 Speaker 1: was pulling directly from an actual interview with a real 205 00:13:02,760 --> 00:13:06,560 Speaker 1: doctor Emily Roberts, the only conclusion we can draw is 206 00:13:06,559 --> 00:13:10,800 Speaker 1: that the AIS fabricating a quote. And I would hate 207 00:13:10,800 --> 00:13:15,200 Speaker 1: to see an AI generated script that purported to feature me. 208 00:13:15,840 --> 00:13:18,520 Speaker 1: And the words that are being spouted off are not 209 00:13:18,679 --> 00:13:22,200 Speaker 1: my own. They're AI generated standing in for me, that 210 00:13:22,520 --> 00:13:26,080 Speaker 1: is really troubling. So I do wonder if chat GPT 211 00:13:26,320 --> 00:13:29,040 Speaker 1: came up with the name Emily Roberts all on its 212 00:13:29,080 --> 00:13:32,480 Speaker 1: own and that's just a coincidence, or if within the 213 00:13:32,600 --> 00:13:35,920 Speaker 1: vast database of information it draws from, in the large 214 00:13:36,000 --> 00:13:39,440 Speaker 1: language model that feeds into chat GPT, there happens to 215 00:13:39,480 --> 00:13:43,360 Speaker 1: be a record of Anne Emily Roberts, who does work 216 00:13:43,400 --> 00:13:47,320 Speaker 1: related to automotive safety. Like maybe that was what statistically 217 00:13:47,360 --> 00:13:50,960 Speaker 1: guided chat GPT to choose that name. I don't know, 218 00:13:51,160 --> 00:13:54,479 Speaker 1: So it's either a coincidence or it's a stone's throwaway 219 00:13:54,840 --> 00:13:59,360 Speaker 1: from impersonation, at least in my opinion. Okay, we're just 220 00:13:59,400 --> 00:14:02,480 Speaker 1: scratching surface here. We've got a lot more to talk 221 00:14:02,520 --> 00:14:06,520 Speaker 1: about as far as this AI generated episode goes. But 222 00:14:06,720 --> 00:14:09,200 Speaker 1: before we get to that, let's take a quick break 223 00:14:09,200 --> 00:14:22,520 Speaker 1: to thank our sponsors. We're back with our AI generated 224 00:14:22,640 --> 00:14:27,400 Speaker 1: technology podcast episode and the issues that I found when 225 00:14:27,480 --> 00:14:31,480 Speaker 1: I did a deeper dive. So, the fictional doctor Emily 226 00:14:31,600 --> 00:14:36,280 Speaker 1: Roberts explains that early prototypes of airbags used compressed air. 227 00:14:36,680 --> 00:14:39,920 Speaker 1: This is true. It is correct at least Hettrick's patent 228 00:14:40,040 --> 00:14:44,560 Speaker 1: refers to quote an air accumulator or reservoir end quote 229 00:14:44,600 --> 00:14:47,200 Speaker 1: that would hold air under pressure, and then there would 230 00:14:47,200 --> 00:14:50,800 Speaker 1: be a valve that would connect this reservoir to deflated 231 00:14:50,880 --> 00:14:54,280 Speaker 1: air bags housed somewhere in the vehicle, such as in 232 00:14:54,400 --> 00:14:58,800 Speaker 1: the steering column and the dashboard. In Hettrick's design, there 233 00:14:58,880 --> 00:15:02,440 Speaker 1: is inside the scent a weight attached to a spring, 234 00:15:02,640 --> 00:15:06,440 Speaker 1: and this controls the valve. So normally the valve's position 235 00:15:06,520 --> 00:15:09,840 Speaker 1: remains firmly closed because the spring, when it's in its 236 00:15:09,960 --> 00:15:14,560 Speaker 1: normal uncompressed state, holds the weight in place. And on 237 00:15:14,600 --> 00:15:17,200 Speaker 1: the other end of the weight is the connection to 238 00:15:17,400 --> 00:15:19,880 Speaker 1: this valve, which is kind of like a plug. Imagine 239 00:15:19,880 --> 00:15:22,520 Speaker 1: a plug in a bathtub, right, and it's connected to 240 00:15:22,560 --> 00:15:27,200 Speaker 1: this thing. When a vehicle were to stop very suddenly, 241 00:15:27,240 --> 00:15:29,840 Speaker 1: such as in the case as a head on collision, 242 00:15:30,200 --> 00:15:34,040 Speaker 1: the momentum the inertia really would mean that the weight 243 00:15:34,080 --> 00:15:37,720 Speaker 1: would continue to move forward even as the vehicle itself stopped. 244 00:15:37,880 --> 00:15:41,040 Speaker 1: Just as people in the car would continue to move forward, right, 245 00:15:41,360 --> 00:15:44,200 Speaker 1: the weight would move forward. This would compress the spring 246 00:15:44,360 --> 00:15:47,440 Speaker 1: that otherwise holds the weight in place, and thus the 247 00:15:47,440 --> 00:15:50,960 Speaker 1: weight would pull the plug forward this valve, and it 248 00:15:50,960 --> 00:15:55,440 Speaker 1: would open the valve between the reservoir of air and 249 00:15:55,480 --> 00:15:58,280 Speaker 1: the deflated air bag. So the idea is the compressed 250 00:15:58,320 --> 00:16:00,960 Speaker 1: air would then rush up of that's reservoir and into 251 00:16:01,000 --> 00:16:03,560 Speaker 1: the area of low pressure, which is the deflated airbag 252 00:16:03,600 --> 00:16:05,880 Speaker 1: in other words, and thus inflate it. So this would 253 00:16:05,920 --> 00:16:10,440 Speaker 1: be a purely mechanical system. There was no actual censor 254 00:16:10,760 --> 00:16:13,760 Speaker 1: to speak of, at least not an electronic one. Momentum 255 00:16:13,760 --> 00:16:16,360 Speaker 1: itself would determine if the weight would open the valve 256 00:16:16,480 --> 00:16:18,960 Speaker 1: or if everything would stay in place. Now you might 257 00:16:19,000 --> 00:16:21,680 Speaker 1: be thinking to yourself, huh, that sounds like it's, you know, 258 00:16:22,320 --> 00:16:26,480 Speaker 1: possible for airbags to deploy if you just stop the 259 00:16:26,480 --> 00:16:30,120 Speaker 1: car really fast, right, like not a collision, but you 260 00:16:30,280 --> 00:16:35,360 Speaker 1: breaked really hard. And yeah, I thought that too. And 261 00:16:35,400 --> 00:16:38,320 Speaker 1: my guess is that is a possibility and of a 262 00:16:38,520 --> 00:16:42,440 Speaker 1: definite drawback on this particular design. At any rate, Hettrick's 263 00:16:42,440 --> 00:16:45,440 Speaker 1: invention didn't really get any traction. It was, as the 264 00:16:45,560 --> 00:16:50,560 Speaker 1: fictitious Emily Roberts points out, not reliable enough, so Hettrick 265 00:16:50,640 --> 00:16:54,000 Speaker 1: never made any money off this particular idea. It would 266 00:16:54,040 --> 00:16:57,680 Speaker 1: serve as kind of a launching point for other inventors, 267 00:16:57,880 --> 00:17:03,120 Speaker 1: but it was not sufficient to actually spur the creation 268 00:17:03,200 --> 00:17:06,719 Speaker 1: of airbags. But I do have another note about that 269 00:17:06,800 --> 00:17:09,920 Speaker 1: nineteen fifty one discrepancy. If you remember, before the break 270 00:17:09,960 --> 00:17:14,159 Speaker 1: I mentioned the AI episode said that Hettrick filed for 271 00:17:14,200 --> 00:17:16,800 Speaker 1: the patent in nineteen fifty one. He didn't. He filed 272 00:17:16,840 --> 00:17:19,159 Speaker 1: for the patent in nineteen fifty two. But there was 273 00:17:19,200 --> 00:17:22,960 Speaker 1: someone else who did file a patent for airbags in 274 00:17:23,040 --> 00:17:26,159 Speaker 1: nineteen fifty one, a whole year before Hettrick did it. 275 00:17:26,400 --> 00:17:31,080 Speaker 1: On October sixth, nineteen fifty one, Walter Linderer applied for 276 00:17:31,200 --> 00:17:34,560 Speaker 1: a patent for his quote device for protecting people in 277 00:17:34,680 --> 00:17:38,200 Speaker 1: vehicles from injuries in the event of a collision end quote. 278 00:17:38,320 --> 00:17:41,320 Speaker 1: Now he filed for this patent in Germany, not the 279 00:17:41,440 --> 00:17:44,240 Speaker 1: United States, so it wasn't the US Patent office, it 280 00:17:44,320 --> 00:17:47,400 Speaker 1: was the German one. That might be why some sources 281 00:17:47,480 --> 00:17:51,119 Speaker 1: cite Hettrick as being the inventor of airbags. It's that 282 00:17:51,160 --> 00:17:54,879 Speaker 1: good old American bias, like it doesn't count unless Americans 283 00:17:54,920 --> 00:17:58,879 Speaker 1: did it. But like Hettrick's design, Linderer proposed using compressed 284 00:17:58,960 --> 00:18:01,520 Speaker 1: air to inflate an air bag in the event of 285 00:18:01,560 --> 00:18:05,399 Speaker 1: an accident. Now Lenderers design allowed for either mechanical or 286 00:18:05,520 --> 00:18:09,920 Speaker 1: electrical control of the valve that would either allow air 287 00:18:09,960 --> 00:18:12,199 Speaker 1: to rush out of the compression chamber or keep it 288 00:18:12,280 --> 00:18:16,240 Speaker 1: stoppered up. Interestingly, in Lenderer's invention, there was the potential 289 00:18:16,280 --> 00:18:19,600 Speaker 1: for manual control, so in other words, the driver could 290 00:18:19,640 --> 00:18:22,159 Speaker 1: throw a lever that would open the valve and deploy 291 00:18:22,200 --> 00:18:25,640 Speaker 1: the airbags. I don't know when this would ever be useful, 292 00:18:25,840 --> 00:18:28,400 Speaker 1: because you know, when you know how fast a car 293 00:18:28,440 --> 00:18:33,199 Speaker 1: accident happens, and how comparatively slowly we humans react to 294 00:18:33,320 --> 00:18:36,480 Speaker 1: this sort of stuff. I don't know that there's any 295 00:18:36,560 --> 00:18:40,919 Speaker 1: scenario where someone would have the time to switch a 296 00:18:41,119 --> 00:18:45,360 Speaker 1: manual lever to deploy airbags and have it be useful. 297 00:18:45,560 --> 00:18:47,800 Speaker 1: If anything, I would think such a lever would actually 298 00:18:47,840 --> 00:18:51,000 Speaker 1: make it more dangerous to operate a car, because what 299 00:18:51,080 --> 00:18:53,040 Speaker 1: if you threw the switch by accident in the middle 300 00:18:53,080 --> 00:18:56,040 Speaker 1: of a trip, and then suddenly airbags deployed. That would 301 00:18:56,080 --> 00:19:00,160 Speaker 1: not be good. But Lenderer also allowed for automatic deploy 302 00:19:00,520 --> 00:19:05,120 Speaker 1: and his proposal was not this mechanical system that Hettrick 303 00:19:05,200 --> 00:19:10,000 Speaker 1: had had thought up. His was an electric contact, actually 304 00:19:10,040 --> 00:19:13,480 Speaker 1: a pair of contacts that would be built into and 305 00:19:14,240 --> 00:19:17,480 Speaker 1: behind the front bumper of the vehicles, so normally the 306 00:19:17,520 --> 00:19:21,000 Speaker 1: two points of contact would not touch. There'd be a 307 00:19:21,040 --> 00:19:23,520 Speaker 1: gap between them. But if the car were to collide 308 00:19:23,560 --> 00:19:27,399 Speaker 1: with something, then the pressure on the bumper would push 309 00:19:27,480 --> 00:19:30,240 Speaker 1: back and cause the contacts to touch one another and 310 00:19:30,280 --> 00:19:34,719 Speaker 1: thus close a circuit, which Linderer saw as drawing power 311 00:19:34,760 --> 00:19:39,399 Speaker 1: from the car's battery to operate the valve electronically and 312 00:19:39,440 --> 00:19:42,920 Speaker 1: to have it open. So it was a rudimentary crash detector, 313 00:19:42,920 --> 00:19:45,240 Speaker 1: in other words, and one that wouldn't be prone to 314 00:19:45,280 --> 00:19:48,119 Speaker 1: the issue of deploying after just a hard breaking you 315 00:19:48,160 --> 00:19:51,959 Speaker 1: would have to have contact with something. While Lenderer would 316 00:19:52,119 --> 00:19:55,280 Speaker 1: file for his patent in Germany a year before Hettrick 317 00:19:55,320 --> 00:19:58,639 Speaker 1: would do the same in the United States, Hettrick actually 318 00:19:58,680 --> 00:20:02,879 Speaker 1: got his patent first. He was awarded his patent about 319 00:20:02,960 --> 00:20:06,240 Speaker 1: three months before Lenderer would get his in Germany. So 320 00:20:06,400 --> 00:20:10,120 Speaker 1: maybe that's why some people think Hettrick is the inventor 321 00:20:10,200 --> 00:20:13,960 Speaker 1: of the airbag, because he got his patent first, even 322 00:20:14,040 --> 00:20:18,520 Speaker 1: though Lenderer filed his a year earlier. I think all 323 00:20:18,560 --> 00:20:21,000 Speaker 1: of this is interesting. I think that this is stuff 324 00:20:21,000 --> 00:20:23,920 Speaker 1: that should be included in any episode about the history 325 00:20:23,920 --> 00:20:26,639 Speaker 1: of airbags, but sadly we don't get any of that 326 00:20:26,920 --> 00:20:29,879 Speaker 1: in the AI generated episode, and rather we seem to 327 00:20:29,880 --> 00:20:32,719 Speaker 1: get a case where the host is conflating two different 328 00:20:32,840 --> 00:20:36,639 Speaker 1: inventors and the invention date. Now I'm not done with 329 00:20:36,680 --> 00:20:39,919 Speaker 1: our fictional historian yet. Doctor Roberts then goes on to 330 00:20:39,920 --> 00:20:43,600 Speaker 1: say that the technology becoming more efficient and reliable and 331 00:20:43,640 --> 00:20:47,119 Speaker 1: going to solid state is what led to the industry 332 00:20:47,200 --> 00:20:52,840 Speaker 1: in general and consumers accepting it. But that's not really true. 333 00:20:53,359 --> 00:20:56,040 Speaker 1: The solid state thing, I don't think really plays into 334 00:20:56,080 --> 00:20:59,680 Speaker 1: it that much, and Chad GPT doesn't seem to factor 335 00:20:59,760 --> 00:21:04,359 Speaker 1: in to account things like the political landscape and the 336 00:21:04,400 --> 00:21:11,560 Speaker 1: automotive industry's flex in US politics, and thus the reason 337 00:21:11,600 --> 00:21:15,680 Speaker 1: why airbags took so long to become mandated safety features 338 00:21:15,680 --> 00:21:18,840 Speaker 1: within vehicles. So there's another cat we got to talk 339 00:21:18,880 --> 00:21:21,639 Speaker 1: about who is left out of the AI version of 340 00:21:21,720 --> 00:21:25,760 Speaker 1: the episode, and his name is Alan K. Bred, an 341 00:21:25,800 --> 00:21:29,639 Speaker 1: engineer and overachiever if ever there was one. So. Breed 342 00:21:29,880 --> 00:21:33,119 Speaker 1: was born in nineteen twenty seven, and by the nineteen 343 00:21:33,200 --> 00:21:36,080 Speaker 1: sixties he had done more than most folks doing an 344 00:21:36,280 --> 00:21:40,360 Speaker 1: entire lifetime. He rushed through a four year college program 345 00:21:40,440 --> 00:21:43,639 Speaker 1: in just two years. He did that by attending different 346 00:21:43,760 --> 00:21:47,879 Speaker 1: lectures simultaneously, and he managed that by leaving a tape 347 00:21:47,880 --> 00:21:50,640 Speaker 1: recorder in one lecture Hall, while he would actually attend 348 00:21:50,880 --> 00:21:55,760 Speaker 1: another lecture. So this guy was driven and ambitious anyway. 349 00:21:55,800 --> 00:21:59,720 Speaker 1: By the late nineteen sixties, he was tackling automotive safety 350 00:21:59,760 --> 00:22:03,360 Speaker 1: and airbag design. The mechanical and electrical switch proposals by 351 00:22:03,359 --> 00:22:06,439 Speaker 1: Hettrick and Lenderer were kind of a non starter. They 352 00:22:06,480 --> 00:22:09,919 Speaker 1: weren't reliable, and anyway, compressed air would not inflate an 353 00:22:09,960 --> 00:22:14,160 Speaker 1: airbag fast enough to provide a cushion for a passenger 354 00:22:14,200 --> 00:22:15,840 Speaker 1: in the event of a crash. It would still be 355 00:22:15,880 --> 00:22:19,440 Speaker 1: inflating long after the passenger had already completed their full 356 00:22:19,520 --> 00:22:21,639 Speaker 1: range of motion in the wake of a crash. So 357 00:22:21,800 --> 00:22:25,639 Speaker 1: Breed innovated both on the crash detection part of the 358 00:22:25,680 --> 00:22:29,320 Speaker 1: problem as well as the airbag inflation bet Now this 359 00:22:29,359 --> 00:22:34,040 Speaker 1: is actually going to overlap into our second fictional guest, 360 00:22:34,320 --> 00:22:38,879 Speaker 1: the supposed automotive safety engineer named James Lee. So James 361 00:22:38,960 --> 00:22:41,560 Speaker 1: Lee is a pretty darn common name, and there were 362 00:22:41,640 --> 00:22:44,679 Speaker 1: lots of hits when I started searching for this to 363 00:22:44,720 --> 00:22:47,879 Speaker 1: see if maybe James Lee was actually a real person 364 00:22:47,920 --> 00:22:50,959 Speaker 1: that was being quoted here. One James Lee I found 365 00:22:51,119 --> 00:22:54,960 Speaker 1: is an engineering manager at a tire company, So you know, 366 00:22:55,680 --> 00:23:00,920 Speaker 1: general automotive industry area. There another served as an in 367 00:23:01,080 --> 00:23:05,120 Speaker 1: turn for automation controls over at TESLA, So again someone 368 00:23:05,160 --> 00:23:08,199 Speaker 1: who's working at least within the industry, though not in 369 00:23:08,240 --> 00:23:11,879 Speaker 1: the same capacity as our supposed guest expert. There's also 370 00:23:12,119 --> 00:23:15,640 Speaker 1: an associate professor at a tech university who specializes in 371 00:23:15,680 --> 00:23:19,879 Speaker 1: manufacturing and mechanical engineering, but not specifically in safety. So 372 00:23:20,160 --> 00:23:22,960 Speaker 1: maybe Chad GBT looked around for a likely name for 373 00:23:23,000 --> 00:23:26,360 Speaker 1: a safety expert and found some instances of various James 374 00:23:26,440 --> 00:23:30,560 Speaker 1: Lees working in jobs relatively in the same industry and 375 00:23:31,080 --> 00:23:34,280 Speaker 1: use that name for this episode. Or maybe again it's 376 00:23:34,520 --> 00:23:37,760 Speaker 1: just a coincidence. Either way, I feel confident saying that, 377 00:23:38,200 --> 00:23:41,399 Speaker 1: like Emily Roberts, there's no real James Lee who is 378 00:23:41,480 --> 00:23:45,359 Speaker 1: being quoted here. Instead, we have an invented expert and 379 00:23:45,480 --> 00:23:50,560 Speaker 1: invented quotations. So back to Breed and his improvements in 380 00:23:50,600 --> 00:23:54,760 Speaker 1: airbag design in the sixties, So on the crash detection front, 381 00:23:54,920 --> 00:23:57,880 Speaker 1: Breed proposed a new kind of electronic sensor that folks 382 00:23:57,920 --> 00:24:01,479 Speaker 1: would refer to as a ball in tube device. So, 383 00:24:01,520 --> 00:24:05,040 Speaker 1: what the heck is a ball in tube? Well, some 384 00:24:05,119 --> 00:24:08,600 Speaker 1: of it is just kind of self explanatory. You got yourself. 385 00:24:08,800 --> 00:24:11,560 Speaker 1: A small ball made of something like steel, something that's 386 00:24:11,600 --> 00:24:14,840 Speaker 1: you know, affected by magnets, and this ball is nestled 387 00:24:14,880 --> 00:24:18,720 Speaker 1: inside a tube and the tube is not affected by magnets. 388 00:24:18,720 --> 00:24:21,399 Speaker 1: It's made out of something like plastic. And there is 389 00:24:21,440 --> 00:24:24,199 Speaker 1: a permanent magnet that holds this ball in place at 390 00:24:24,320 --> 00:24:27,159 Speaker 1: one end of the tube. But with enough force, the 391 00:24:27,240 --> 00:24:29,960 Speaker 1: ball can be dislodged. It can break free of this 392 00:24:30,080 --> 00:24:32,800 Speaker 1: magnetic hold, and at that point it rolls down the 393 00:24:32,840 --> 00:24:36,960 Speaker 1: tube and ultimately it hits an electrical trigger that closes 394 00:24:37,040 --> 00:24:39,679 Speaker 1: a circuit, so a contact in other words, or it 395 00:24:39,720 --> 00:24:43,359 Speaker 1: pushes two points of contact to touch one another. So 396 00:24:43,440 --> 00:24:47,480 Speaker 1: in some ways Breed's design married what Hetrick and Linderer 397 00:24:47,520 --> 00:24:51,199 Speaker 1: were both working on independently. Now, this is not a 398 00:24:51,240 --> 00:24:55,399 Speaker 1: solid state electronic device. Solid state electronics are based on 399 00:24:55,440 --> 00:24:59,959 Speaker 1: semiconductors like microchips essentially, So this is an electro mechanical 400 00:25:00,560 --> 00:25:05,200 Speaker 1: switch that detects crashes and it and variations of this 401 00:25:05,320 --> 00:25:08,320 Speaker 1: would be used in air bags for decades. So modern 402 00:25:08,400 --> 00:25:12,080 Speaker 1: cars do have lots of semiconductor microchips, tons of them, 403 00:25:12,200 --> 00:25:15,720 Speaker 1: like more than one hundred in your typical automobile, and 404 00:25:15,800 --> 00:25:21,159 Speaker 1: these include everything from the chips that connect to entertainment systems, 405 00:25:21,359 --> 00:25:25,720 Speaker 1: to navigation systems, to yes safety systems, including air bags. 406 00:25:25,960 --> 00:25:31,800 Speaker 1: But the sensors themselves work on mechanical principles involving inertia. Right. 407 00:25:31,880 --> 00:25:34,800 Speaker 1: We're not talking about a solid state chip that detects 408 00:25:34,800 --> 00:25:39,159 Speaker 1: a crash, not usually anyway. We're talking about this electro 409 00:25:39,240 --> 00:25:42,000 Speaker 1: mechanical switch. There are a couple of different variations of this, 410 00:25:42,200 --> 00:25:45,080 Speaker 1: so some of them aren't ball and tube systems. Some 411 00:25:45,240 --> 00:25:48,720 Speaker 1: use a cam like a little swinging Think of it 412 00:25:48,800 --> 00:25:51,520 Speaker 1: like a lever. It's a swinging lever and it can 413 00:25:51,560 --> 00:25:54,600 Speaker 1: freely swing, but in the event of an impact, it'll 414 00:25:55,000 --> 00:25:58,240 Speaker 1: swing hard enough to push two contacts together and close 415 00:25:58,280 --> 00:26:02,000 Speaker 1: a circuit, thus sending the command deploy airbags. Other sensors 416 00:26:02,119 --> 00:26:05,840 Speaker 1: use a coiled spring. So think of like one of 417 00:26:05,880 --> 00:26:10,119 Speaker 1: those party favors, like the classic party favor, where you 418 00:26:10,160 --> 00:26:15,280 Speaker 1: blow into it and it unwraps this long paper tube 419 00:26:15,600 --> 00:26:18,680 Speaker 1: that otherwise is coiled up. Same sort of thing, right, 420 00:26:18,800 --> 00:26:21,879 Speaker 1: except at the very end of this coil, like in 421 00:26:21,920 --> 00:26:25,119 Speaker 1: the interior part of the coil, is a contact, and 422 00:26:25,160 --> 00:26:29,320 Speaker 1: when the coil is uncoiled, that contact can touch a 423 00:26:29,359 --> 00:26:32,600 Speaker 1: second contact that completes a circuit. So again, normally it's 424 00:26:32,600 --> 00:26:37,320 Speaker 1: coiled up, but with a sudden, sudden stop, like an impact, 425 00:26:37,640 --> 00:26:41,760 Speaker 1: the momentum will force this to uncoil and thus the 426 00:26:41,840 --> 00:26:45,040 Speaker 1: contacts come into touch with one another and boom, you 427 00:26:45,080 --> 00:26:48,280 Speaker 1: get your air bag deployed. So again, interesting stuff. But 428 00:26:48,359 --> 00:26:51,960 Speaker 1: our safety engineer James Lee and our historian doctor Emily Roberts, 429 00:26:52,000 --> 00:26:54,919 Speaker 1: they don't touch on any of this, right, But I 430 00:26:54,960 --> 00:27:00,000 Speaker 1: feel like these things understanding how these sensors work is 431 00:27:00,000 --> 00:27:02,720 Speaker 1: it's really cool and interesting, but it's not at all 432 00:27:02,760 --> 00:27:07,560 Speaker 1: covered in the episode. Also, solid state electronics and automobiles 433 00:27:07,800 --> 00:27:10,440 Speaker 1: would become a thing well before the nineteen eighties, which 434 00:27:10,520 --> 00:27:13,000 Speaker 1: is what doctor Emily Roberts said was what led to 435 00:27:13,119 --> 00:27:16,800 Speaker 1: the adoption of airbags, right, was that solid state crash 436 00:27:16,880 --> 00:27:21,920 Speaker 1: detectors were really one of the driving components. I'm using 437 00:27:21,960 --> 00:27:23,960 Speaker 1: a lot of puns here, but the AI one did too, 438 00:27:24,040 --> 00:27:26,639 Speaker 1: They just didn't acknowledge it, but one of the driving 439 00:27:26,680 --> 00:27:29,679 Speaker 1: components that would lead to the adoption of airbags. Solid 440 00:27:29,680 --> 00:27:33,320 Speaker 1: state electronics were already a thing well before the nineteen eighties. 441 00:27:33,320 --> 00:27:36,800 Speaker 1: And nineteen seventy two article in Popular Electronics by John D. 442 00:27:37,040 --> 00:27:40,520 Speaker 1: Drummond details a list of features in upcoming vehicles that 443 00:27:40,560 --> 00:27:43,800 Speaker 1: would rely upon solid state electronics and microchips. This was 444 00:27:43,880 --> 00:27:46,600 Speaker 1: kind of like the article talking about the automobiles of 445 00:27:46,600 --> 00:27:50,359 Speaker 1: the future. Drummond also glumbly pointed out in that article 446 00:27:50,520 --> 00:27:53,360 Speaker 1: that vehicles that were most likely to feature this kind 447 00:27:53,359 --> 00:27:56,240 Speaker 1: of cutting edge technology would actually come from places like 448 00:27:56,320 --> 00:27:59,600 Speaker 1: Japan and Germany rather than the United States, because the 449 00:27:59,640 --> 00:28:04,639 Speaker 1: American automotive industry did not have strong collaborative relationships with 450 00:28:04,800 --> 00:28:08,600 Speaker 1: electronics companies and semiconductor companies the same way that foreign 451 00:28:08,640 --> 00:28:12,000 Speaker 1: car companies did. Like he said, these car companies had 452 00:28:12,080 --> 00:28:16,680 Speaker 1: reached out and proactively created these relationships with other companies 453 00:28:16,680 --> 00:28:20,119 Speaker 1: that together could collaborate on these kinds of systems. But 454 00:28:20,200 --> 00:28:23,960 Speaker 1: in America that wasn't happening. That Detroit was very insular 455 00:28:24,400 --> 00:28:27,320 Speaker 1: and they weren't eager to reach out to other companies 456 00:28:27,359 --> 00:28:29,680 Speaker 1: and collaborate on this kind of stuff. At any rate, 457 00:28:29,760 --> 00:28:32,480 Speaker 1: Solid state components and vehicles were in place well before 458 00:28:32,520 --> 00:28:34,840 Speaker 1: we would get up to the nineteen eighties. Even in 459 00:28:35,280 --> 00:28:38,560 Speaker 1: safety features now, the AI, James Lee does reveal that 460 00:28:38,640 --> 00:28:41,760 Speaker 1: compressed air wasn't the way to go, and Breed agreed 461 00:28:41,800 --> 00:28:44,960 Speaker 1: with that back in the nineteen sixties. Lee cites a 462 00:28:45,040 --> 00:28:48,480 Speaker 1: chemical compound called sodium azide, and sure enough, that's a 463 00:28:48,520 --> 00:28:53,480 Speaker 1: compound found in a lot of airbags, particularly older airbags. However, 464 00:28:53,680 --> 00:28:56,880 Speaker 1: the industry has moved away from sodium azide in large 465 00:28:56,880 --> 00:28:59,200 Speaker 1: part due to the fact that the compound is both 466 00:28:59,280 --> 00:29:03,920 Speaker 1: highly react and extremely toxic. We'll talk more about how 467 00:29:03,960 --> 00:29:07,520 Speaker 1: airbags actually work in just a moment, but first let's 468 00:29:07,520 --> 00:29:19,920 Speaker 1: take another quick break. Okay, before the break, I was 469 00:29:19,960 --> 00:29:24,000 Speaker 1: alluding to the fact that airbags don't rely on compressed air. 470 00:29:24,520 --> 00:29:28,719 Speaker 1: That compressed air just isn't a fast enough agent to 471 00:29:28,920 --> 00:29:32,320 Speaker 1: inflate an airbag in the event of a crash. So 472 00:29:32,440 --> 00:29:36,720 Speaker 1: let's talk about what happens with the chemicals that are 473 00:29:36,840 --> 00:29:41,520 Speaker 1: in airbags, such as sodium azide. Well, sodium azide isn't 474 00:29:41,520 --> 00:29:43,960 Speaker 1: the only one at the party in these older airbags. 475 00:29:44,040 --> 00:29:46,560 Speaker 1: There are other chemicals that are in there too. Some 476 00:29:46,720 --> 00:29:50,200 Speaker 1: are meant to accelerate a chemical reaction. So just like 477 00:29:50,960 --> 00:29:53,280 Speaker 1: you know, it's like pouring gasoline on a fire, and 478 00:29:53,360 --> 00:29:57,280 Speaker 1: it really some of the chemicals are meant to be 479 00:29:57,400 --> 00:30:01,160 Speaker 1: the ignition for the chemical reaction, and so literally they're 480 00:30:01,200 --> 00:30:07,680 Speaker 1: the most reactive element inside this mix, and they will 481 00:30:07,720 --> 00:30:11,640 Speaker 1: then initiate a larger chemical reaction that will be responsible 482 00:30:11,640 --> 00:30:16,480 Speaker 1: for inflating the bag. Some of the chemical agents inside 483 00:30:16,520 --> 00:30:20,480 Speaker 1: air bags are there to combine with the byproducts of 484 00:30:20,520 --> 00:30:23,640 Speaker 1: the chemical reaction that inflates the bag in order to 485 00:30:23,680 --> 00:30:26,840 Speaker 1: make those byproducts safer for humans to be around them. 486 00:30:27,120 --> 00:30:30,200 Speaker 1: But essentially what happens is you have an electric charge 487 00:30:30,240 --> 00:30:34,520 Speaker 1: which ignites the ignition chemicals inside the air bag. They 488 00:30:34,560 --> 00:30:38,280 Speaker 1: in turn cause a further reaction, which is essentially an explosion, 489 00:30:38,560 --> 00:30:43,960 Speaker 1: and the solid compound rapidly converts to new gaseous components, 490 00:30:44,120 --> 00:30:47,880 Speaker 1: inflating the air bag in the process. So with sodium azide, 491 00:30:48,040 --> 00:30:53,040 Speaker 1: the explosion creates sodium and nitrogen. Now, sodium is pretty 492 00:30:53,200 --> 00:30:58,200 Speaker 1: darn dangerous to humans, is highly reactive stuff, so typically 493 00:30:58,200 --> 00:31:02,160 Speaker 1: airbags would also contain something like silica and potassium nitrate. 494 00:31:02,520 --> 00:31:06,520 Speaker 1: This would bind with that sodium and thus render it inactive, 495 00:31:06,680 --> 00:31:09,760 Speaker 1: so it wouldn't be dangerous to be around. Heaven help 496 00:31:09,840 --> 00:31:12,440 Speaker 1: you if that sodium were to come into contact with 497 00:31:12,560 --> 00:31:16,720 Speaker 1: water first, that would be disastrous. So this reaction happens 498 00:31:16,880 --> 00:31:20,960 Speaker 1: very very quickly. It is explosive, so in less than 499 00:31:21,000 --> 00:31:25,120 Speaker 1: thirty milliseconds you'll have an inflated airbag because of this 500 00:31:25,240 --> 00:31:28,560 Speaker 1: with the nitrogen gas that's generated filling it. Now, like 501 00:31:28,640 --> 00:31:31,840 Speaker 1: I said, a lot of modern airbags actually use other 502 00:31:31,920 --> 00:31:35,520 Speaker 1: chemicals to get the same effect as sodium azide. These days, 503 00:31:35,960 --> 00:31:39,080 Speaker 1: sodium azide just isn't as common, at least not to 504 00:31:39,120 --> 00:31:43,920 Speaker 1: the degree that the AI created James Lee would suggest. 505 00:31:44,280 --> 00:31:46,560 Speaker 1: So this is another bone I have to pick with 506 00:31:46,640 --> 00:31:51,800 Speaker 1: the AI generated episode. It gives a definitive answer saying 507 00:31:51,840 --> 00:31:56,320 Speaker 1: sodium azide is like the most common airbag ingredient today, 508 00:31:56,680 --> 00:32:01,040 Speaker 1: and that's not necessarily true. It was true, but it's 509 00:32:01,120 --> 00:32:04,840 Speaker 1: not really true now, Like a lot of the industry 510 00:32:04,880 --> 00:32:07,680 Speaker 1: has moved away from sodiumazite for various reasons, mostly to 511 00:32:07,720 --> 00:32:11,040 Speaker 1: do with safety. So yeah, I take issue with that 512 00:32:11,080 --> 00:32:13,959 Speaker 1: particular part of the AI generated episode too. All right, 513 00:32:14,000 --> 00:32:16,280 Speaker 1: let's get back to another claim that was made by 514 00:32:16,680 --> 00:32:21,480 Speaker 1: doctor Emily Roberts, the automotive industry historian, And she had 515 00:32:21,480 --> 00:32:25,320 Speaker 1: said that the industry was quick to adopt airbags once 516 00:32:25,560 --> 00:32:29,120 Speaker 1: they became efficient and less bulky, But that's not what 517 00:32:29,240 --> 00:32:32,160 Speaker 1: was holding up the automotive industry really. What was holding 518 00:32:32,160 --> 00:32:35,160 Speaker 1: them up was that adding in components that don't have 519 00:32:35,440 --> 00:32:40,120 Speaker 1: a strong customer demand attached to them and ones that 520 00:32:40,160 --> 00:32:43,960 Speaker 1: are not mandated by some sort of government means that 521 00:32:44,360 --> 00:32:47,040 Speaker 1: you don't put it into your car because it just 522 00:32:47,120 --> 00:32:49,560 Speaker 1: adds to the cost of production and you're not going 523 00:32:49,640 --> 00:32:55,520 Speaker 1: to realize a higher profit from using that particular technology. 524 00:32:55,680 --> 00:32:58,000 Speaker 1: If no one's asking for it and the government's not 525 00:32:58,040 --> 00:33:00,640 Speaker 1: requiring it, it's not going to go in because it 526 00:33:00,720 --> 00:33:03,080 Speaker 1: just doesn't make sense from a business perspective. That was 527 00:33:03,120 --> 00:33:06,280 Speaker 1: the view of the automotive industry. The automotive industry in 528 00:33:06,320 --> 00:33:10,600 Speaker 1: general in the United States long resisted lots of attempts 529 00:33:10,640 --> 00:33:14,480 Speaker 1: to impose safety regulations because it would mean things would 530 00:33:14,520 --> 00:33:17,080 Speaker 1: get more expensive and it would be hard to pass 531 00:33:17,120 --> 00:33:19,680 Speaker 1: that expense on to the customer and plus make even 532 00:33:19,720 --> 00:33:24,400 Speaker 1: more profit because customers weren't super excited about safety systems. 533 00:33:24,440 --> 00:33:28,160 Speaker 1: That wasn't really what was being marketed to drivers in 534 00:33:28,200 --> 00:33:31,479 Speaker 1: the nineteen fifties and nineteen sixties. They wanted cars that 535 00:33:31,520 --> 00:33:35,200 Speaker 1: looked fun, fast. They didn't really care if they were safe. 536 00:33:35,600 --> 00:33:38,800 Speaker 1: Why incorporate safety devices in your car if you didn't 537 00:33:38,840 --> 00:33:41,480 Speaker 1: have to, and if market research didn't show that there 538 00:33:41,560 --> 00:33:43,360 Speaker 1: was some sort of demand for it, it's not like 539 00:33:43,720 --> 00:33:46,600 Speaker 1: you could use it in order to sell more units, 540 00:33:47,080 --> 00:33:48,800 Speaker 1: and if the government didn't give a fig one way 541 00:33:48,880 --> 00:33:50,760 Speaker 1: or the other, you might as well just save yourself 542 00:33:50,760 --> 00:33:53,479 Speaker 1: the money and leave it out of the design entirely. 543 00:33:53,880 --> 00:33:57,040 Speaker 1: That's the real reason airbags would not become a standard 544 00:33:57,040 --> 00:34:01,200 Speaker 1: feature for decades, not that folks aren't trying to change that. 545 00:34:01,800 --> 00:34:04,560 Speaker 1: In the nineteen sixties, a lawyer who would become a 546 00:34:04,600 --> 00:34:08,719 Speaker 1: future politician, Ralph Nader. He wrote a book titled Unsafe 547 00:34:08,719 --> 00:34:13,160 Speaker 1: at Any Speed, The designed in Dangers of the American Automobile, 548 00:34:13,680 --> 00:34:16,680 Speaker 1: and it really raked the American auto industry over the 549 00:34:16,760 --> 00:34:20,560 Speaker 1: coals for ignoring safety concerns and failing to include proper 550 00:34:20,600 --> 00:34:25,160 Speaker 1: safety technologies, including basic stuff like seat belts, and generally 551 00:34:25,560 --> 00:34:29,360 Speaker 1: the automotive industry contributing to an environment in which thousands 552 00:34:29,360 --> 00:34:33,600 Speaker 1: of people were experiencing terrible tragedies, like thousands of people 553 00:34:33,640 --> 00:34:37,400 Speaker 1: were dying or getting really badly injured because of traffic 554 00:34:37,480 --> 00:34:41,440 Speaker 1: accidents and a lack of safety equipment. To say that 555 00:34:41,640 --> 00:34:46,560 Speaker 1: this book was influential is putting it lightly. It was 556 00:34:46,600 --> 00:34:49,799 Speaker 1: incredibly popular. It was one of the most popular nonfiction 557 00:34:49,960 --> 00:34:52,719 Speaker 1: books on the bestseller list in nineteen sixty six, and 558 00:34:52,760 --> 00:34:55,640 Speaker 1: it really pressured the US government to create the Department 559 00:34:55,680 --> 00:34:59,640 Speaker 1: of Transportation that year. Later on the US government would 560 00:34:59,640 --> 00:35:04,280 Speaker 1: also create the National Highway Traffic Safety Administration in HTSA. 561 00:35:04,320 --> 00:35:07,520 Speaker 1: That would be in nineteen seventy. Nator's book singled out 562 00:35:07,560 --> 00:35:10,880 Speaker 1: a specific make and model, which was the Chevrolet Corvare, 563 00:35:11,120 --> 00:35:14,920 Speaker 1: but his criticisms extended to the automotive industry in general 564 00:35:15,000 --> 00:35:18,440 Speaker 1: in America. Now, some states already had laws in place 565 00:35:18,480 --> 00:35:22,000 Speaker 1: that required auto manufacturers to include seat belts and vehicles 566 00:35:22,040 --> 00:35:25,080 Speaker 1: sold in that state, but it wasn't a US federal 567 00:35:25,160 --> 00:35:30,080 Speaker 1: requirement until nineteen sixty eight. A requirement to include airbags 568 00:35:30,280 --> 00:35:34,719 Speaker 1: would wait quite a bit longer. It would not officially 569 00:35:34,800 --> 00:35:39,200 Speaker 1: become a mandated requirement for passenger vehicles and light trucks 570 00:35:39,360 --> 00:35:44,520 Speaker 1: until September one, nineteen ninety eight. That was actually seven 571 00:35:44,600 --> 00:35:47,799 Speaker 1: years after the respective law had already been passed. It 572 00:35:47,840 --> 00:35:50,560 Speaker 1: was a nineteen ninety one law, but didn't go into 573 00:35:50,640 --> 00:35:54,120 Speaker 1: effect until ninety eight. Now, that long gap was partly 574 00:35:54,160 --> 00:35:57,080 Speaker 1: there to give manufacturers time to take all the necessary 575 00:35:57,120 --> 00:36:00,840 Speaker 1: steps to incorporate airbags into vehicle design. But yeah, the 576 00:36:00,880 --> 00:36:04,839 Speaker 1: airbag requirement in the US is actually fairly recent. One 577 00:36:04,880 --> 00:36:07,200 Speaker 1: reason it took so long for airbags to become a 578 00:36:07,280 --> 00:36:10,319 Speaker 1: required safety feature in the US is that initially they 579 00:36:10,320 --> 00:36:13,799 Speaker 1: were being viewed as an alternative to seat belts, not 580 00:36:13,960 --> 00:36:18,359 Speaker 1: an augmented complementary safety feature. You weren't necessarily talking about 581 00:36:18,400 --> 00:36:21,520 Speaker 1: a car that had both seat belts and air bags, 582 00:36:21,680 --> 00:36:26,240 Speaker 1: but rather seat belts or air bags, and auto manufacturers 583 00:36:26,239 --> 00:36:28,640 Speaker 1: were able to convince the US government that seat belts 584 00:36:28,680 --> 00:36:31,719 Speaker 1: were a much better safety feature than airbags alone, which 585 00:36:31,760 --> 00:36:35,359 Speaker 1: you know, not necessarily false, that is true, like it'd 586 00:36:35,400 --> 00:36:37,400 Speaker 1: be best to have both of them, But that's not 587 00:36:37,440 --> 00:36:40,200 Speaker 1: what the auto manufacturers were focusing on. They were saying, well, 588 00:36:40,280 --> 00:36:41,880 Speaker 1: if you're going to require us to have these, at 589 00:36:41,960 --> 00:36:44,800 Speaker 1: least make us have the ones that are more effective 590 00:36:44,840 --> 00:36:48,080 Speaker 1: and also significantly cheaper, though they didn't say that part 591 00:36:48,080 --> 00:36:51,160 Speaker 1: out loud so much so the matter was tabled for 592 00:36:51,200 --> 00:36:54,680 Speaker 1: a couple of decades. Airbags just weren't a requirement, but 593 00:36:54,719 --> 00:36:57,840 Speaker 1: seat belts were. Now. A few car companies did experiment 594 00:36:57,880 --> 00:37:01,680 Speaker 1: with including airbags before it would become required. In the 595 00:37:01,719 --> 00:37:05,399 Speaker 1: early nineteen seventies, both GM and Ford would test air 596 00:37:05,440 --> 00:37:09,240 Speaker 1: bags in fleets of vehicles, but these were not consumer vehicles. 597 00:37:09,239 --> 00:37:12,600 Speaker 1: They weren't passenger vehicles that just any old schmoe could 598 00:37:12,640 --> 00:37:16,000 Speaker 1: go buy off a lot. The first passenger vehicle that 599 00:37:16,120 --> 00:37:18,840 Speaker 1: was sold to the general public that had air bags 600 00:37:18,840 --> 00:37:23,399 Speaker 1: as an option was the nineteen seventy three Oldsmobile Toronado. 601 00:37:24,120 --> 00:37:28,640 Speaker 1: GM called it the Air Cushion Restraint System. The Toronado 602 00:37:29,080 --> 00:37:31,680 Speaker 1: was a luxury car, so it was not the sort 603 00:37:31,719 --> 00:37:34,320 Speaker 1: of thing that, again, an average car buyer would be 604 00:37:34,360 --> 00:37:37,440 Speaker 1: in the market for. Like it was way above the 605 00:37:37,520 --> 00:37:40,040 Speaker 1: average person's price range for a vehicle, and it would 606 00:37:40,040 --> 00:37:42,200 Speaker 1: take quite some time for air bags to become a 607 00:37:42,239 --> 00:37:46,560 Speaker 1: more common option in other makes and models. Safety, however, 608 00:37:46,680 --> 00:37:50,720 Speaker 1: remained a big concern for Ralph Nader, largely because while 609 00:37:51,040 --> 00:37:54,760 Speaker 1: the nineteen sixty eight law required manufacturers to include seat belts, 610 00:37:54,920 --> 00:37:58,920 Speaker 1: there was no federal law requiring drivers and passengers to 611 00:37:59,040 --> 00:38:02,840 Speaker 1: actually use use them, and apparently like twenty five percent 612 00:38:02,920 --> 00:38:05,560 Speaker 1: of folks in cars actually bothered to buckle up and 613 00:38:05,600 --> 00:38:09,280 Speaker 1: everyone else didn't. To this day, there is no federal 614 00:38:09,400 --> 00:38:13,680 Speaker 1: law in the United States that requires seat belt use. However, 615 00:38:13,840 --> 00:38:17,279 Speaker 1: every state in the US, with the exception of New Hampshire, 616 00:38:17,480 --> 00:38:22,160 Speaker 1: requires everybody in a vehicle to wear seat belts, and 617 00:38:22,320 --> 00:38:25,840 Speaker 1: every state also has laws about seat belts for children. 618 00:38:26,080 --> 00:38:29,200 Speaker 1: New Hampshire. If you're an adult, you don't have to 619 00:38:29,239 --> 00:38:32,480 Speaker 1: wear a seat belt, but that's that's the only exception. 620 00:38:32,800 --> 00:38:35,720 Speaker 1: Penalties for not wearing seat belts vary from state to state. 621 00:38:36,120 --> 00:38:39,560 Speaker 1: Some states, the police are allowed to stop someone if 622 00:38:39,600 --> 00:38:41,960 Speaker 1: they see that that person's not wearing a seat belt, 623 00:38:42,120 --> 00:38:45,800 Speaker 1: but in other states that's not a big enough crime 624 00:38:46,160 --> 00:38:49,359 Speaker 1: for a cop to be able to justify pulling someone over. 625 00:38:49,640 --> 00:38:52,200 Speaker 1: Now it is a secondary offense, So if the person's 626 00:38:52,239 --> 00:38:55,080 Speaker 1: also doing something else, like failing to signal or whatever, 627 00:38:55,200 --> 00:38:57,960 Speaker 1: something that is enough for a cop to pull them over, 628 00:38:58,120 --> 00:39:00,359 Speaker 1: they can also get cited for not wearing a seat 629 00:39:00,360 --> 00:39:03,200 Speaker 1: belt in that particular instance, but they're not allowed to 630 00:39:03,200 --> 00:39:05,120 Speaker 1: be pulled over just for not wearing a seat belt 631 00:39:05,160 --> 00:39:09,000 Speaker 1: in those particular states. Gradually, more car manufacturers began to 632 00:39:09,040 --> 00:39:13,560 Speaker 1: offer vehicles with airbags because it turned out that over time, 633 00:39:14,200 --> 00:39:19,000 Speaker 1: the desires of the American buying public were changing and 634 00:39:19,040 --> 00:39:23,600 Speaker 1: safety was becoming a bigger concern. So airbags became a 635 00:39:23,640 --> 00:39:26,239 Speaker 1: selling point for certain types of cars and the folks 636 00:39:26,280 --> 00:39:29,319 Speaker 1: who typically shop for those types of cars. So there 637 00:39:29,320 --> 00:39:32,600 Speaker 1: were market forces guiding the automotive industry to start to 638 00:39:32,680 --> 00:39:35,560 Speaker 1: embrace the air bag, at least for certain models, but 639 00:39:35,600 --> 00:39:38,280 Speaker 1: the industry as a whole continued to resist any efforts 640 00:39:38,320 --> 00:39:41,719 Speaker 1: to make air bags a required safety feature. Again, this 641 00:39:41,840 --> 00:39:44,080 Speaker 1: was largely because air bags would add to the cost 642 00:39:44,080 --> 00:39:47,040 Speaker 1: of manufacturing, and there was a concern that customers wouldn't 643 00:39:47,080 --> 00:39:49,239 Speaker 1: be so happy to foot the extra expense, plus you know, 644 00:39:49,280 --> 00:39:51,799 Speaker 1: whatever the profit margin the car company had in mind 645 00:39:51,840 --> 00:39:55,360 Speaker 1: on top of that. So the airbags issue got battered 646 00:39:55,400 --> 00:39:59,000 Speaker 1: around back and forth legally due to various political and 647 00:39:59,160 --> 00:40:03,160 Speaker 1: industrial reasons. So in nineteen seventy seven, under the Jimmy 648 00:40:03,200 --> 00:40:07,600 Speaker 1: Carter presidency, a rule would have required all car manufacturers 649 00:40:07,640 --> 00:40:11,279 Speaker 1: to start including airbags or automatic seat belts starting in 650 00:40:11,280 --> 00:40:14,920 Speaker 1: the nineteen eighties. However, in nineteen eighty one, under the 651 00:40:14,960 --> 00:40:20,279 Speaker 1: Reagan administration, as part of a widespread attempt to deregulate everything, 652 00:40:20,760 --> 00:40:23,600 Speaker 1: this rule was struck down before it could go into effect, 653 00:40:23,920 --> 00:40:28,680 Speaker 1: but in nineteen eighty three, the Supreme Court overturned that decision. 654 00:40:29,000 --> 00:40:31,840 Speaker 1: In nineteen eighty four, the US passed an amendment to 655 00:40:32,000 --> 00:40:35,360 Speaker 1: federal motor safety standards that would require all manufacturers to 656 00:40:35,440 --> 00:40:39,840 Speaker 1: include a passive restraint for the driver of any vehicle 657 00:40:40,160 --> 00:40:43,239 Speaker 1: that was made after April first, nineteen eighty nine, so 658 00:40:44,160 --> 00:40:47,479 Speaker 1: that could either be an airbag or an automatic seat belt, 659 00:40:47,480 --> 00:40:50,520 Speaker 1: because again, not everyone would even bother buckling up when 660 00:40:50,520 --> 00:40:53,839 Speaker 1: they would get in, so automatic seat belts took that 661 00:40:53,960 --> 00:40:57,160 Speaker 1: decision out of their hands. It would automatically end up 662 00:40:57,400 --> 00:41:00,600 Speaker 1: buckling someone in. By the way, automatic seat belts were 663 00:41:00,800 --> 00:41:04,240 Speaker 1: often cited as being less effective than normal seat belts, 664 00:41:04,239 --> 00:41:06,640 Speaker 1: which is a reason why a lot of automotive industry 665 00:41:06,640 --> 00:41:09,880 Speaker 1: experts were arguing against it, although another reason is again 666 00:41:10,000 --> 00:41:12,839 Speaker 1: automatic seat belt systems tend to be more expensive than 667 00:41:12,920 --> 00:41:16,680 Speaker 1: just passive seat belts. The mandatory requirement for airbags really 668 00:41:16,719 --> 00:41:19,600 Speaker 1: did not become a thing until a nineteen ninety one 669 00:41:19,760 --> 00:41:22,440 Speaker 1: revision to US laws, and again that didn't actually go 670 00:41:22,480 --> 00:41:27,239 Speaker 1: into effect until nineteen ninety eight. None of this has 671 00:41:27,280 --> 00:41:31,640 Speaker 1: to do with the technology of airbags or their efficiency. 672 00:41:31,680 --> 00:41:34,919 Speaker 1: It has everything to do with an industry resisting regulations 673 00:41:34,960 --> 00:41:38,239 Speaker 1: in an effort to avoid taking on more costs. So 674 00:41:39,040 --> 00:41:42,839 Speaker 1: I feel the AI generated episode does a disservice by 675 00:41:42,920 --> 00:41:46,040 Speaker 1: painting an oversimplistic picture of what was going on. But 676 00:41:46,160 --> 00:41:49,200 Speaker 1: let's breeze through the rest of this AI generated episode. 677 00:41:49,280 --> 00:41:53,040 Speaker 1: So our next guest expert is Lisa Fernandez. She appears 678 00:41:53,080 --> 00:41:56,600 Speaker 1: to be another fabricated expert. I guess it's possible that 679 00:41:56,800 --> 00:42:01,960 Speaker 1: the former Olympic softball player Lisa Fernandez, who now is 680 00:42:02,040 --> 00:42:05,920 Speaker 1: a coach at UCLA, is also really into air bags, 681 00:42:06,040 --> 00:42:09,919 Speaker 1: but I couldn't find any information supporting that. So this 682 00:42:10,200 --> 00:42:14,080 Speaker 1: presumably fake guest does talk about different kinds of airbags, 683 00:42:14,120 --> 00:42:17,640 Speaker 1: and there certainly are all kinds of different airbags, including 684 00:42:17,680 --> 00:42:21,000 Speaker 1: things like side impact airbags and curtain air bags. So 685 00:42:21,120 --> 00:42:25,399 Speaker 1: this section from a content standpoint is more or less 686 00:42:25,440 --> 00:42:30,359 Speaker 1: fine once you get past again an invented expert. Next up, 687 00:42:30,560 --> 00:42:34,080 Speaker 1: we have another really important bit, which is safety statistics 688 00:42:34,080 --> 00:42:37,160 Speaker 1: and airbags, and this has a supposed expert from the 689 00:42:37,239 --> 00:42:42,680 Speaker 1: National Highway Traffic Safety Administration or NHTSA, a statistician named 690 00:42:42,719 --> 00:42:46,080 Speaker 1: doctor Michael Harris. Now I found a safety claims adjuster 691 00:42:46,360 --> 00:42:49,120 Speaker 1: who goes by Mike Harris. I found a doctor Michael 692 00:42:49,120 --> 00:42:53,000 Speaker 1: Harris who is a biological safety scientist and biomedical engineer 693 00:42:53,000 --> 00:42:57,320 Speaker 1: at NAMSA. There's an Atlanta Braves center fielder named Michael Harris. 694 00:42:57,320 --> 00:43:00,400 Speaker 1: The second, but again we appear to have ours a 695 00:43:00,400 --> 00:43:04,080 Speaker 1: fake expert, although maybe Michael Harris the second can talk 696 00:43:04,120 --> 00:43:07,160 Speaker 1: to Lisa Fernandez. They're both coming from the world of 697 00:43:07,200 --> 00:43:11,480 Speaker 1: sports and apparently getting involved in automotive safety. Again. When 698 00:43:11,560 --> 00:43:17,560 Speaker 1: you have a supposed statistician from an official agency providing 699 00:43:17,680 --> 00:43:21,160 Speaker 1: actual safety figures, but it turns out that person is invented, 700 00:43:21,440 --> 00:43:25,240 Speaker 1: I think that's an enormous ethical problem personally a bulk 701 00:43:25,280 --> 00:43:29,960 Speaker 1: at that Anyway, chat GPT's host says that the NHTSA 702 00:43:30,000 --> 00:43:32,440 Speaker 1: claims thousands of lives have been saved due to the 703 00:43:32,440 --> 00:43:37,200 Speaker 1: introduction of airbags. That checks out. The NHTSA's overview on airbags, 704 00:43:37,200 --> 00:43:40,280 Speaker 1: which is available on the web at the NHTSA website, 705 00:43:40,400 --> 00:43:44,080 Speaker 1: says quote Frontal airbags have saved more than fifty thousand 706 00:43:44,120 --> 00:43:46,799 Speaker 1: lives over a thirty year period end quote. So we're 707 00:43:46,800 --> 00:43:52,360 Speaker 1: good there. Next up, this supposed expert cites that airbags 708 00:43:52,400 --> 00:43:56,680 Speaker 1: contribute to a reduction in fatalities for frontal collisions for 709 00:43:56,760 --> 00:44:00,799 Speaker 1: both drivers and right front passengers. There's a twenty nine 710 00:44:00,800 --> 00:44:04,080 Speaker 1: percent reduction in driver fatalities and a thirty two percent 711 00:44:04,160 --> 00:44:08,000 Speaker 1: reduction in passenger fatalities. Now I was able to track 712 00:44:08,239 --> 00:44:11,600 Speaker 1: these statistics down. They come from a twenty fifteen report, 713 00:44:11,680 --> 00:44:15,160 Speaker 1: and it is from the NHTSA. It is titled Lives 714 00:44:15,280 --> 00:44:19,960 Speaker 1: Saved by Vehicle Safety Technologies and Associated Federal Motor Vehicle 715 00:44:20,040 --> 00:44:24,120 Speaker 1: Safety Standards nineteen sixty to twenty twelve. The information on 716 00:44:24,200 --> 00:44:27,239 Speaker 1: driver fatality reductions is on page one twenty five. The 717 00:44:27,280 --> 00:44:30,160 Speaker 1: one on right front passengers is on page one twenty seven. 718 00:44:30,320 --> 00:44:31,799 Speaker 1: Just in case you want to check up on it 719 00:44:31,800 --> 00:44:36,320 Speaker 1: yourselves now, this expert also cites a statistic saying side 720 00:44:36,320 --> 00:44:40,080 Speaker 1: air bags with head protection reduced driver fatalities by thirty 721 00:44:40,120 --> 00:44:43,600 Speaker 1: seven percent in cars and fifty two percent in SUVs. 722 00:44:44,040 --> 00:44:47,120 Speaker 1: This figure seems to come from a totally different document. 723 00:44:47,200 --> 00:44:49,879 Speaker 1: This one was written in two thousand and seven by 724 00:44:50,080 --> 00:44:55,960 Speaker 1: Anne McCart and Sergei Kirichinko. It's titled Efficacy of side 725 00:44:55,960 --> 00:44:59,160 Speaker 1: air Bags and Reducing driver Deaths in driver side car 726 00:44:59,320 --> 00:45:04,279 Speaker 1: and Suv collisions. So these authors work was published in 727 00:45:04,320 --> 00:45:08,319 Speaker 1: a journal titled Traffic Injury Prevention. This is not an 728 00:45:08,480 --> 00:45:13,960 Speaker 1: NHTSA publication, so this particular bit being attributed to an 729 00:45:14,080 --> 00:45:18,200 Speaker 1: NHTSA statistician is a bit of an issue. However, Traffic 730 00:45:18,320 --> 00:45:22,319 Speaker 1: Injury Prevention is a journal with peer review. I think 731 00:45:22,320 --> 00:45:25,680 Speaker 1: they use a single anonymized peer to review papers. I 732 00:45:25,719 --> 00:45:29,840 Speaker 1: don't know how rigorous the review processes, but it is 733 00:45:29,880 --> 00:45:32,760 Speaker 1: peer reviewed, and I would say that traffic injury Prevention 734 00:45:32,960 --> 00:45:37,200 Speaker 1: is probably a reliable source. It's just not the NHTSA, 735 00:45:37,239 --> 00:45:41,279 Speaker 1: and I would argue the chat GPT generated interview implies otherwise. 736 00:45:41,600 --> 00:45:45,680 Speaker 1: So this information does reflect official studies, though from one 737 00:45:45,719 --> 00:45:48,920 Speaker 1: report that's nearly a decade old, and on top of that, 738 00:45:49,239 --> 00:45:53,920 Speaker 1: this report actually covers information that spans between nineteen sixty 739 00:45:53,960 --> 00:45:56,680 Speaker 1: to twenty twelve, so it's even more dated. And then 740 00:45:56,719 --> 00:45:59,880 Speaker 1: the other piece is from a paper that was published 741 00:45:59,920 --> 00:46:02,760 Speaker 1: in two thousand and seven. I don't have more recent 742 00:46:02,800 --> 00:46:05,359 Speaker 1: reports to pull information from, or at least I didn't 743 00:46:05,440 --> 00:46:08,120 Speaker 1: uncover any while I was researching this, But I have 744 00:46:08,200 --> 00:46:12,080 Speaker 1: to imagine that the statistics are different today. I'm guessing 745 00:46:12,120 --> 00:46:16,080 Speaker 1: they're likely better today than they were when this initial 746 00:46:16,120 --> 00:46:19,520 Speaker 1: report was written in the NHTSA. And that's just a 747 00:46:19,560 --> 00:46:21,960 Speaker 1: guess on my part. I actually don't have the data 748 00:46:21,960 --> 00:46:24,759 Speaker 1: to back it up, but I remain I still have 749 00:46:24,800 --> 00:46:29,240 Speaker 1: ethical concerns about the fact that they invented a statistician 750 00:46:29,360 --> 00:46:34,400 Speaker 1: representing the NHTSA in this AI generated podcast However, at 751 00:46:34,480 --> 00:46:37,960 Speaker 1: least the data that was pulled seems to be more 752 00:46:38,040 --> 00:46:41,719 Speaker 1: or less above board, so that's good. Okay, we still 753 00:46:41,719 --> 00:46:43,880 Speaker 1: have one more expert to go, and I have a 754 00:46:43,880 --> 00:46:46,919 Speaker 1: lot more thoughts to share about this, but we're running along, 755 00:46:46,960 --> 00:46:59,320 Speaker 1: so let's take another quick break and we'll be right back. Okay. 756 00:46:59,360 --> 00:47:05,560 Speaker 1: So final and yes, fictional guest expert is doctor Sarah Kim. 757 00:47:06,000 --> 00:47:08,160 Speaker 1: I looked for a doctor Sarah Kim who is an 758 00:47:08,200 --> 00:47:11,560 Speaker 1: automotive safety expert. I could not find one. There are 759 00:47:11,560 --> 00:47:14,399 Speaker 1: plenty of Sarah Kim's, obviously, but I could not find 760 00:47:14,440 --> 00:47:18,960 Speaker 1: one who would have been an expert that a podcast 761 00:47:19,000 --> 00:47:23,600 Speaker 1: host would cite on an episode about airbag safety. But 762 00:47:24,120 --> 00:47:29,200 Speaker 1: this expert says that AI and advanced sensors will improve 763 00:47:29,239 --> 00:47:33,279 Speaker 1: airbag deployment situations. Right now, you have an airbag micro 764 00:47:33,400 --> 00:47:36,680 Speaker 1: controller like a microcomputer essentially that's in charge of this 765 00:47:36,800 --> 00:47:42,000 Speaker 1: kind of stuff, and your typical airbag microcomputer system, it's 766 00:47:42,040 --> 00:47:45,160 Speaker 1: detecting which seats have occupants, so that way it quote 767 00:47:45,239 --> 00:47:48,360 Speaker 1: unquote knows which airbags should be deployed in the event 768 00:47:48,400 --> 00:47:51,160 Speaker 1: of a crash. So when it detects a crash, it 769 00:47:51,280 --> 00:47:55,799 Speaker 1: sends signals to the appropriate airbags to inflate and then 770 00:47:55,880 --> 00:47:59,640 Speaker 1: other ones don't have to. So there are microcomputers that 771 00:47:59,760 --> 00:48:02,759 Speaker 1: are part of this. This is those solid state electronics 772 00:48:02,760 --> 00:48:05,120 Speaker 1: that doctor Emily Roberts talk to us about at the 773 00:48:05,120 --> 00:48:08,760 Speaker 1: beginning of the episode. But these sensors also can gauge 774 00:48:08,800 --> 00:48:13,520 Speaker 1: the actual impact. So if it determines that it wasn't 775 00:48:13,719 --> 00:48:16,799 Speaker 1: a hard enough stop, like if it wasn't a great 776 00:48:16,920 --> 00:48:20,160 Speaker 1: enough impact, then it can prevent the airbags from going off. 777 00:48:20,520 --> 00:48:23,400 Speaker 1: So if you have like a little fender bender bump, 778 00:48:23,719 --> 00:48:26,840 Speaker 1: then your area airbags aren't going to just spontaneously deploy 779 00:48:27,000 --> 00:48:30,160 Speaker 1: in most cases, and that's good because otherwise it can 780 00:48:30,200 --> 00:48:32,279 Speaker 1: make things much much worse, like the airbags need to 781 00:48:32,280 --> 00:48:37,480 Speaker 1: deploy when the crash is severe, but otherwise probably shouldn't happen. 782 00:48:37,640 --> 00:48:40,040 Speaker 1: Doctor Kim makes a pretty safe bet that this is 783 00:48:40,160 --> 00:48:42,760 Speaker 1: just going to get better in the future, which I feel. 784 00:48:43,480 --> 00:48:45,680 Speaker 1: I feel that that's one of those cop out answers 785 00:48:45,680 --> 00:48:48,520 Speaker 1: which I'm guilty of doing myself. It's kind of like 786 00:48:49,480 --> 00:48:51,520 Speaker 1: I hate doing it now, but man, I used to 787 00:48:51,520 --> 00:48:53,759 Speaker 1: be real bad about this, where I would say, I 788 00:48:53,760 --> 00:48:56,240 Speaker 1: guess we'll have to wait and see, or some variation 789 00:48:56,360 --> 00:48:58,680 Speaker 1: on that. Occasionally I'll still do it, and I hate 790 00:48:58,719 --> 00:49:00,919 Speaker 1: myself every time I do. It's just one of those 791 00:49:00,960 --> 00:49:04,040 Speaker 1: trophy kind of things. Now, the fake doctor Kem also 792 00:49:04,120 --> 00:49:07,880 Speaker 1: says that airbags will protect people outside the car and 793 00:49:07,920 --> 00:49:10,520 Speaker 1: that these are in development. There are cars that are 794 00:49:10,520 --> 00:49:13,840 Speaker 1: going to have external air bags. This is kind of 795 00:49:13,960 --> 00:49:17,920 Speaker 1: true because as far back as twenty twelve, Volvo was 796 00:49:18,640 --> 00:49:22,200 Speaker 1: using an air bag system in which a vehicle detecting 797 00:49:22,239 --> 00:49:24,600 Speaker 1: an impact with a pedestrian or a cyclist would deploy 798 00:49:24,680 --> 00:49:28,000 Speaker 1: an external airbag. Essentially, it was meant to cush in 799 00:49:28,040 --> 00:49:31,960 Speaker 1: a person from hitting the windshield of the Volvo. So 800 00:49:32,320 --> 00:49:36,000 Speaker 1: the Volvo's hood or bonnet, if you're across the pond 801 00:49:36,200 --> 00:49:39,239 Speaker 1: would lift up close to the windshield and an air 802 00:49:39,320 --> 00:49:43,000 Speaker 1: bag would inflate there. And in fact, this feature was 803 00:49:43,320 --> 00:49:46,360 Speaker 1: present in the Volvo V forty. This was not just 804 00:49:46,640 --> 00:49:49,759 Speaker 1: a prototype, it wasn't a concept. It actually made it 805 00:49:49,800 --> 00:49:53,480 Speaker 1: into a production line vehicle. But Volvo did discontinue the 806 00:49:53,560 --> 00:49:56,319 Speaker 1: V forty in twenty nineteen, and as far as I 807 00:49:56,320 --> 00:49:59,320 Speaker 1: can tell, I haven't seen any other vehicles that actually 808 00:49:59,320 --> 00:50:02,000 Speaker 1: have this kind of system in it. But yeah, it 809 00:50:02,040 --> 00:50:07,080 Speaker 1: did exist for one line of cars for a while. Now, overall, 810 00:50:07,280 --> 00:50:13,440 Speaker 1: I think chat GPT made a pretty mediocre tech podcast episode. 811 00:50:14,040 --> 00:50:18,759 Speaker 1: Most of the information wasn't outright wrong or misleading, but 812 00:50:18,840 --> 00:50:22,640 Speaker 1: it was incomplete. It didn't tell the whole story, and 813 00:50:22,680 --> 00:50:26,080 Speaker 1: in some cases it kind of gave an oversimplification and 814 00:50:26,400 --> 00:50:29,960 Speaker 1: misdirect on the story. It also required a lot of 815 00:50:30,120 --> 00:50:33,200 Speaker 1: fact checking on my part, which makes my heart go 816 00:50:33,280 --> 00:50:36,080 Speaker 1: out to any editors out there whose job is to 817 00:50:37,080 --> 00:50:40,319 Speaker 1: read over AI generated articles and make sure that they 818 00:50:40,360 --> 00:50:44,200 Speaker 1: are accurate, that they make sense, because as short as 819 00:50:44,200 --> 00:50:47,200 Speaker 1: that episode was, it took me hours to go through 820 00:50:47,320 --> 00:50:50,360 Speaker 1: all of the different statements and to check them and 821 00:50:50,400 --> 00:50:54,120 Speaker 1: make sure that they were accurate or see what was missing. 822 00:50:54,480 --> 00:50:58,360 Speaker 1: The creation of fake experts is incredibly concerning to me. 823 00:50:59,200 --> 00:51:03,360 Speaker 1: I don't know why why chat gpt made that choice, Like, 824 00:51:03,440 --> 00:51:07,680 Speaker 1: why was it necessary to invent experts to whom the 825 00:51:07,719 --> 00:51:11,480 Speaker 1: host could talk? Why not just create a pair of 826 00:51:11,880 --> 00:51:14,719 Speaker 1: podcast hosts. If you want to have it be a dialogue, 827 00:51:15,040 --> 00:51:18,359 Speaker 1: make two fictional hosts who talk to each other and 828 00:51:18,400 --> 00:51:22,680 Speaker 1: they just cite information, right, They just say the NHTSA 829 00:51:22,840 --> 00:51:26,680 Speaker 1: says that, blah blah blah. Why create people out of 830 00:51:26,760 --> 00:51:30,560 Speaker 1: whole cloth to stand in as experts? That is what's 831 00:51:30,600 --> 00:51:32,520 Speaker 1: going to get you into trouble. You might remember the 832 00:51:32,560 --> 00:51:36,120 Speaker 1: story there was a lawyer who presented arguments that turned 833 00:51:36,160 --> 00:51:39,680 Speaker 1: out to be AI generated, and the problem was the 834 00:51:39,760 --> 00:51:44,720 Speaker 1: AI created case histories that didn't exist, So the lawyer 835 00:51:44,800 --> 00:51:48,080 Speaker 1: was citing precedents that wasn't even real. And when the 836 00:51:48,200 --> 00:51:50,760 Speaker 1: judge reviewed it and said, hey, hang on a minute, 837 00:51:50,920 --> 00:51:55,759 Speaker 1: none of the cases that you're referencing in this report exist, 838 00:51:56,120 --> 00:52:00,480 Speaker 1: the lawyer got into really deep trouble and write so 839 00:52:01,080 --> 00:52:04,560 Speaker 1: like they were just relying upon AI being accurate and 840 00:52:04,640 --> 00:52:07,680 Speaker 1: not just making stuff up. We see even in this 841 00:52:07,800 --> 00:52:11,040 Speaker 1: simple example, that's just not the case. Like, I don't 842 00:52:11,120 --> 00:52:14,280 Speaker 1: like the idea of anyone just making up an expert. 843 00:52:14,360 --> 00:52:16,880 Speaker 1: I could do that if I wanted to in episodes. 844 00:52:16,960 --> 00:52:19,920 Speaker 1: It would make the episodes a lot easier if I 845 00:52:20,080 --> 00:52:24,960 Speaker 1: just invented experts who gave me supposedly reliable information. But 846 00:52:25,040 --> 00:52:29,160 Speaker 1: it's not ethical. And what's with offuscating the sources of 847 00:52:29,200 --> 00:52:31,319 Speaker 1: the information in the first place. Why couldn't it just 848 00:52:31,400 --> 00:52:34,800 Speaker 1: tell me where the information originally came from. I probably 849 00:52:34,840 --> 00:52:37,960 Speaker 1: should have followed up and asked more specific questions about that, 850 00:52:38,360 --> 00:52:41,239 Speaker 1: just to see did it ultimately pull this information from 851 00:52:41,280 --> 00:52:44,920 Speaker 1: the primary source. Did it pull the info from the 852 00:52:45,120 --> 00:52:48,880 Speaker 1: actual report where the information first appeared, or did it 853 00:52:48,920 --> 00:52:53,160 Speaker 1: pull it from some blog post or article that cited 854 00:52:53,800 --> 00:52:56,960 Speaker 1: that initial report. My guess is the second, but I 855 00:52:56,960 --> 00:52:59,000 Speaker 1: don't know for sure because I didn't follow up, So 856 00:52:59,040 --> 00:53:03,160 Speaker 1: I can't reilly fault chat GPT for that yet because 857 00:53:03,239 --> 00:53:06,320 Speaker 1: I did not take the steps needed to determine whether 858 00:53:06,480 --> 00:53:10,040 Speaker 1: or not it was being lazy quote unquote, but off 859 00:53:10,080 --> 00:53:12,440 Speaker 1: these skating the sources just seems weird. Like when I 860 00:53:12,480 --> 00:53:15,080 Speaker 1: asked it, what sources did you use? That seems like 861 00:53:15,080 --> 00:53:17,560 Speaker 1: a pretty straightforward question that it would be able to 862 00:53:17,960 --> 00:53:20,239 Speaker 1: rattle off, like, oh, it was pulled from this and 863 00:53:20,280 --> 00:53:23,080 Speaker 1: this and this and this, But that was just not 864 00:53:23,160 --> 00:53:25,279 Speaker 1: the case. If you're the kind of person who wants 865 00:53:25,360 --> 00:53:28,600 Speaker 1: to be sure that the stuff you're consuming is accurate, 866 00:53:29,040 --> 00:53:33,000 Speaker 1: it makes it really much more difficult. If the AI 867 00:53:33,320 --> 00:53:36,799 Speaker 1: bot can't give you the sources it's pulling from, it 868 00:53:36,800 --> 00:53:39,800 Speaker 1: doesn't make it impossible. Clearly I did it, like I 869 00:53:39,800 --> 00:53:42,160 Speaker 1: did all the research. It is a lot more work 870 00:53:42,440 --> 00:53:45,040 Speaker 1: on your part, to the point where again, I wonder, 871 00:53:45,080 --> 00:53:48,319 Speaker 1: what's the point in having the AI synthesize everything? If 872 00:53:48,320 --> 00:53:51,640 Speaker 1: you can't be sure that what you're getting is really 873 00:53:51,680 --> 00:53:55,919 Speaker 1: accurate and meaningful. Yeah, I feel like this experience has 874 00:53:56,080 --> 00:53:59,960 Speaker 1: reaffirmed a lot of my concerns around generative AI in general. 875 00:54:00,280 --> 00:54:02,319 Speaker 1: But that's not to say that it's always going to 876 00:54:02,360 --> 00:54:06,320 Speaker 1: be this obtuse. I hope that there will be greater 877 00:54:06,400 --> 00:54:09,920 Speaker 1: transparency in the future so that it becomes easier to 878 00:54:10,000 --> 00:54:13,800 Speaker 1: make sure that the stuff you're being presented is accurate. 879 00:54:13,960 --> 00:54:18,000 Speaker 1: This also will depend upon AI training itself on data 880 00:54:18,200 --> 00:54:22,600 Speaker 1: that didn't get generated by other less capable versions of AI. 881 00:54:22,840 --> 00:54:25,440 Speaker 1: We talked about that in an earlier episode about how 882 00:54:25,920 --> 00:54:29,520 Speaker 1: that kind of process can lead to model collapse, where 883 00:54:29,520 --> 00:54:33,400 Speaker 1: a large language model essentially crumbles in on itself because 884 00:54:33,400 --> 00:54:37,520 Speaker 1: it's been trained on AI generated material. And the more 885 00:54:37,600 --> 00:54:41,480 Speaker 1: that happens, the poorer the quality is of the output. Right, 886 00:54:41,600 --> 00:54:45,480 Speaker 1: garbage and garbage out essentially. But yeah, that was an 887 00:54:45,560 --> 00:54:49,960 Speaker 1: AI generated episode of a technology podcast and then me 888 00:54:50,120 --> 00:54:53,319 Speaker 1: eviscerating it. I do think that generative AI can be 889 00:54:53,400 --> 00:54:56,160 Speaker 1: great for thought starters, like I think if I had 890 00:54:56,239 --> 00:54:59,880 Speaker 1: used this just in general to make a podcast about 891 00:55:00,040 --> 00:55:02,560 Speaker 1: whatever tech topic, and then I went through and actually 892 00:55:02,600 --> 00:55:06,680 Speaker 1: researched everything, it could help me structure an episode. It 893 00:55:06,719 --> 00:55:09,320 Speaker 1: could give me direction on the things that are really 894 00:55:09,360 --> 00:55:13,640 Speaker 1: important that I think is valuable. What I worry though, 895 00:55:13,840 --> 00:55:18,080 Speaker 1: is that a lot of people and organizations are skipping that. 896 00:55:18,120 --> 00:55:20,759 Speaker 1: They're not really worried about that step. They're just using 897 00:55:20,880 --> 00:55:23,799 Speaker 1: AI to generate as much junk as they can in 898 00:55:23,880 --> 00:55:28,000 Speaker 1: order to monetize it. To just blast out a massive 899 00:55:28,000 --> 00:55:31,600 Speaker 1: amount of content on the internet, post ads against it 900 00:55:32,000 --> 00:55:35,399 Speaker 1: and hope that the money rolls in from SEO. That 901 00:55:35,680 --> 00:55:40,840 Speaker 1: is dreadful. Like I've never liked content farms ever. I 902 00:55:40,920 --> 00:55:44,080 Speaker 1: always hated whenever house Stuff Works would do stuff that 903 00:55:44,800 --> 00:55:48,800 Speaker 1: kind of started to tread a little close to content farms. 904 00:55:49,120 --> 00:55:52,480 Speaker 1: I'm talking about, like lots of image galleries and slideshows 905 00:55:52,520 --> 00:55:56,719 Speaker 1: and puzzles or quizzes that would have lots of page 906 00:55:56,800 --> 00:56:00,040 Speaker 1: views added to them. I always felt gross about that. 907 00:56:00,360 --> 00:56:03,120 Speaker 1: It wasn't my call, but I really didn't like it. 908 00:56:03,560 --> 00:56:09,080 Speaker 1: I much preferred writing high quality articles whenever possible. But yeah, 909 00:56:09,200 --> 00:56:13,320 Speaker 1: that's not always the world we live in. And content farms, 910 00:56:13,360 --> 00:56:17,560 Speaker 1: while they don't provide much of value in my opinion, 911 00:56:18,520 --> 00:56:22,680 Speaker 1: they're getting worse through AI generation. So that's it. That 912 00:56:22,880 --> 00:56:26,680 Speaker 1: was this episode being written by AI kinda. I hope 913 00:56:26,760 --> 00:56:29,040 Speaker 1: you enjoyed it. I hope you learned something. I hope, 914 00:56:29,080 --> 00:56:32,560 Speaker 1: if anything else, this convinces you to not rely too 915 00:56:32,600 --> 00:56:35,920 Speaker 1: heavily on generative AI for stuff that you have to create, 916 00:56:36,320 --> 00:56:39,319 Speaker 1: using it as again a thought starter or a way 917 00:56:39,360 --> 00:56:42,560 Speaker 1: to guide you when you're structuring things. I think that's 918 00:56:42,600 --> 00:56:46,760 Speaker 1: perfectly cromulent, as the Simpsons would say, but I don't 919 00:56:46,840 --> 00:56:49,520 Speaker 1: think it's a good idea to just lean on that 920 00:56:49,600 --> 00:56:52,160 Speaker 1: to do your work for you. It doesn't do a 921 00:56:52,200 --> 00:56:56,040 Speaker 1: good enough job. It misses important and interesting things that 922 00:56:56,800 --> 00:57:00,680 Speaker 1: I think are absolutely crucial to having a deeper understanding 923 00:57:00,680 --> 00:57:02,759 Speaker 1: of the subject matter. And it might tell you some 924 00:57:02,800 --> 00:57:06,319 Speaker 1: wrong stuff, like when somebody filed a patent, so keep 925 00:57:06,360 --> 00:57:09,080 Speaker 1: that in mind. That's it for this episode of tech Stuff. 926 00:57:09,520 --> 00:57:12,279 Speaker 1: I hope all of you are well, and I will 927 00:57:12,280 --> 00:57:21,560 Speaker 1: talk to you again really soon. Tech Stuff is an 928 00:57:21,600 --> 00:57:27,120 Speaker 1: iHeartRadio production. For more podcasts from iHeartRadio, visit the iHeartRadio app, 929 00:57:27,240 --> 00:57:30,400 Speaker 1: Apple Podcasts, or wherever you listen to your favorite shows.