1 00:00:04,440 --> 00:00:12,399 Speaker 1: Welcome to Tech Stuff, a production from iHeartRadio. Hey there, 2 00:00:12,400 --> 00:00:16,000 Speaker 1: and welcome to tech Stuff. I'm your host, Jonathan Strickland. 3 00:00:16,079 --> 00:00:19,160 Speaker 1: I'm an executive producer with iHeartRadio. And how the tech 4 00:00:19,200 --> 00:00:22,600 Speaker 1: are you. It's time for some tech news for June 5 00:00:22,760 --> 00:00:26,159 Speaker 1: of first, twenty twenty three. How did we get to 6 00:00:26,360 --> 00:00:32,120 Speaker 1: June already? Yikes? All right, it's another AI heavy news 7 00:00:32,159 --> 00:00:35,440 Speaker 1: episode because you know that's what's going on out there. 8 00:00:35,800 --> 00:00:39,560 Speaker 1: You might recall that on Tuesday I talked about Stephen Schwartz, 9 00:00:39,720 --> 00:00:43,800 Speaker 1: the lawyer who submitted a filing in a legal case 10 00:00:43,800 --> 00:00:48,320 Speaker 1: that contained false information courtesy of chat GPT. He didn't 11 00:00:48,320 --> 00:00:52,400 Speaker 1: do it on purpose. Chat GPT just gave him information 12 00:00:52,479 --> 00:00:56,160 Speaker 1: that was not true. Schwartz had used chat GPT as 13 00:00:56,200 --> 00:00:58,680 Speaker 1: part of his legal research in a case, and the 14 00:00:58,760 --> 00:01:03,560 Speaker 1: chat bot invented some cases that never actually existed as precedents. 15 00:01:03,920 --> 00:01:07,319 Speaker 1: So Schwartz did actually think to ask chat gpt if 16 00:01:07,360 --> 00:01:10,680 Speaker 1: the cases were real, Like there's a there's an exchange 17 00:01:10,680 --> 00:01:12,520 Speaker 1: where he said, hey, is that a real case? In 18 00:01:12,600 --> 00:01:17,840 Speaker 1: chat GPT is like yeah, yeah toats toad's real. So 19 00:01:17,920 --> 00:01:20,959 Speaker 1: it turns out you just can't trust a chatbot. This, 20 00:01:21,080 --> 00:01:23,480 Speaker 1: by the way, is why. I'm really concerned about these 21 00:01:23,560 --> 00:01:29,640 Speaker 1: chatbots being incorporated directly into things like web search, you know, 22 00:01:29,680 --> 00:01:32,800 Speaker 1: with Microsoft and Google both rushing to do that. I 23 00:01:32,800 --> 00:01:36,360 Speaker 1: think it's a mistake because of multiple reasons, one of 24 00:01:36,400 --> 00:01:41,000 Speaker 1: which is this tendency toward hallucinations. Now Schwartz is awaiting 25 00:01:41,000 --> 00:01:43,640 Speaker 1: a hearing that will happen on June eighth that will 26 00:01:43,640 --> 00:01:47,200 Speaker 1: determine what, if any, sanctions he will face as a 27 00:01:47,280 --> 00:01:53,840 Speaker 1: result of this goof them up. Meanwhile, in Texas, a 28 00:01:54,960 --> 00:01:58,360 Speaker 1: judge named Brantley Starr has made it clear that he 29 00:01:58,400 --> 00:02:02,320 Speaker 1: will not abide any AI chat bought shenanigans in his court. 30 00:02:02,720 --> 00:02:05,640 Speaker 1: He said that attorneys in his courtroom must promise that 31 00:02:05,760 --> 00:02:09,840 Speaker 1: quote no portion of the filing was drafted by generative 32 00:02:09,960 --> 00:02:14,359 Speaker 1: artificial intelligence end quote, or if any part of the 33 00:02:14,440 --> 00:02:20,320 Speaker 1: filing did involve generative AI in any respect, a human 34 00:02:20,400 --> 00:02:24,160 Speaker 1: being must have checked the information and verified it to 35 00:02:24,200 --> 00:02:29,399 Speaker 1: be true and accurate. This covers pretty much anything lawyers 36 00:02:29,440 --> 00:02:32,519 Speaker 1: would submit to the court, and I think it's an 37 00:02:32,600 --> 00:02:36,839 Speaker 1: excellent idea. As we saw with Schwartz, AI just it's 38 00:02:36,960 --> 00:02:40,800 Speaker 1: not trustworthy. It can make stuff up in some cases, 39 00:02:41,000 --> 00:02:44,160 Speaker 1: and honestly, this step is good for everyone in the 40 00:02:44,240 --> 00:02:47,760 Speaker 1: legal system. In the long run, I wouldn't be surprised 41 00:02:48,000 --> 00:02:52,640 Speaker 1: to see other judges follow suit. Open AI, meanwhile, is 42 00:02:52,720 --> 00:02:57,800 Speaker 1: trying to address this troubling problem of AI hallucinations. So yes, 43 00:02:58,000 --> 00:03:01,280 Speaker 1: in case you forgot what halluci nations mean with regard 44 00:03:01,320 --> 00:03:04,480 Speaker 1: to AI, it just means an incident in which AI 45 00:03:04,720 --> 00:03:08,600 Speaker 1: invents information, such as the fake legal cases cited in 46 00:03:08,639 --> 00:03:13,120 Speaker 1: Schwartz's situation. And it's not that AI is a pathological 47 00:03:13,200 --> 00:03:16,280 Speaker 1: liar or has some sort of motivation to give us 48 00:03:16,320 --> 00:03:20,760 Speaker 1: the wrong information. It's more like, when this AI gets 49 00:03:20,800 --> 00:03:23,280 Speaker 1: into a situation where it does not have all the 50 00:03:23,320 --> 00:03:28,240 Speaker 1: relevant information, sometimes it just makes stuff up in the 51 00:03:28,320 --> 00:03:32,760 Speaker 1: absence of reliable info. To me, it's kind of like 52 00:03:32,800 --> 00:03:35,840 Speaker 1: if you've ever had a friend who just seems incapable 53 00:03:35,920 --> 00:03:38,680 Speaker 1: of using the words I don't know the answer to that, 54 00:03:39,640 --> 00:03:43,760 Speaker 1: then you probably feel like this is a very familiar situation, right. 55 00:03:44,800 --> 00:03:47,280 Speaker 1: Just think of someone who, rather than say I don't know, 56 00:03:47,720 --> 00:03:51,240 Speaker 1: that's interesting, I don't know, they say, oh, it's probably 57 00:03:51,320 --> 00:03:54,400 Speaker 1: because or maybe they don't even you know, try to aculivocate. 58 00:03:54,480 --> 00:03:57,720 Speaker 1: They just outright say something they think is probably true 59 00:03:58,040 --> 00:04:00,160 Speaker 1: and they don't know. One way or the other. It's 60 00:04:00,200 --> 00:04:04,440 Speaker 1: kind of what open ai says is happening with chat GPT, 61 00:04:04,600 --> 00:04:07,000 Speaker 1: actually what a lot of AI experts say are happening 62 00:04:07,000 --> 00:04:10,760 Speaker 1: with generative AI in general. And so now the company 63 00:04:10,800 --> 00:04:14,400 Speaker 1: is saying it's going to revisit how AI works toward 64 00:04:14,880 --> 00:04:19,560 Speaker 1: creating an answer. So right now, the model apparently follows 65 00:04:19,560 --> 00:04:23,840 Speaker 1: a process called outcome supervision, in which the goal is 66 00:04:23,920 --> 00:04:26,440 Speaker 1: just to get the final answer. It doesn't really matter 67 00:04:26,480 --> 00:04:30,279 Speaker 1: what pathway you took to get there. The ends justify 68 00:04:30,279 --> 00:04:34,719 Speaker 1: the means. In other words, so the outcome supervision is 69 00:04:34,839 --> 00:04:37,600 Speaker 1: just the AI gets a reward if the answer it 70 00:04:37,640 --> 00:04:40,640 Speaker 1: provides at the end of the day is correct. The 71 00:04:40,720 --> 00:04:44,520 Speaker 1: problem is that when AI makes a mistakes, say early 72 00:04:44,600 --> 00:04:46,880 Speaker 1: on in the process, this can have a much larger 73 00:04:46,920 --> 00:04:49,760 Speaker 1: effect further on in the process. Like if you've ever 74 00:04:49,839 --> 00:04:53,719 Speaker 1: put something together and you made a mistake early, you 75 00:04:53,800 --> 00:04:55,640 Speaker 1: might realize that by the time you're getting toward the 76 00:04:55,720 --> 00:04:59,000 Speaker 1: end that small mistake has created a situation much further 77 00:04:59,040 --> 00:05:02,400 Speaker 1: in the process. That is a huge problem. Well, so 78 00:05:02,640 --> 00:05:05,080 Speaker 1: is the same with AI, And so open ai is 79 00:05:05,080 --> 00:05:10,279 Speaker 1: saying they're looking at changing over to process supervision in 80 00:05:10,320 --> 00:05:14,320 Speaker 1: which you know, reward stages for the AI occur throughout 81 00:05:14,440 --> 00:05:19,520 Speaker 1: the reasoning process, so this thought is supposed to less 82 00:05:19,839 --> 00:05:24,200 Speaker 1: think that you know, AI would reward itself every step 83 00:05:24,240 --> 00:05:26,720 Speaker 1: along the way as it made the right choices, and 84 00:05:26,760 --> 00:05:31,520 Speaker 1: thus would reduce the possibility of making a mistake. Further 85 00:05:31,640 --> 00:05:36,800 Speaker 1: down the line, critics argue that it may not make 86 00:05:36,880 --> 00:05:41,200 Speaker 1: any difference at all to the amount of misinformation or 87 00:05:41,240 --> 00:05:45,320 Speaker 1: just fake information that is generated by AI. It might 88 00:05:45,360 --> 00:05:48,720 Speaker 1: not matter what process it uses, but rather what's more 89 00:05:48,760 --> 00:05:53,160 Speaker 1: important is that AI operates with a lack of transparency, 90 00:05:53,240 --> 00:05:56,039 Speaker 1: so it can be really hard to pinpoint where a 91 00:05:56,120 --> 00:06:00,800 Speaker 1: problem starts because you can't actually see what the processes. 92 00:06:00,880 --> 00:06:02,600 Speaker 1: And if you can't see what the process is, it's 93 00:06:02,720 --> 00:06:06,640 Speaker 1: very hard to diagnose where the problem is popping up. 94 00:06:07,480 --> 00:06:09,920 Speaker 1: And so the critics worry that this change in method 95 00:06:09,920 --> 00:06:14,440 Speaker 1: won't actually solve the problem of AI creating incorrect responses 96 00:06:14,480 --> 00:06:19,320 Speaker 1: and misinformation. Over in Italy, a senator named Marco Lombardo 97 00:06:19,680 --> 00:06:23,719 Speaker 1: stood before Parliament and delivered a speech about Italy's agreements 98 00:06:23,800 --> 00:06:28,440 Speaker 1: with Switzerland, and that does not sound particularly techi except 99 00:06:28,920 --> 00:06:32,320 Speaker 1: at the conclusion of the speech, Lombardo revealed that the 100 00:06:32,360 --> 00:06:34,760 Speaker 1: speech he read out was not written by a human 101 00:06:34,800 --> 00:06:39,000 Speaker 1: being instead it was generated by AI, and further, Lombardo 102 00:06:39,080 --> 00:06:42,560 Speaker 1: said he did this in order to prompt a larger 103 00:06:42,680 --> 00:06:48,160 Speaker 1: conversation about AI and to really consider what it can 104 00:06:48,240 --> 00:06:52,839 Speaker 1: do and the potential consequences that can occur if people 105 00:06:53,360 --> 00:06:57,039 Speaker 1: misuse it or if the AI does not perform as expected. 106 00:06:57,520 --> 00:07:00,719 Speaker 1: Italy has been one of the more proactive country to 107 00:07:00,960 --> 00:07:05,160 Speaker 1: consider AI and to be critical of AI. Previously, Italy 108 00:07:05,279 --> 00:07:09,720 Speaker 1: banned chat gpt, although only temporarily, and did so because 109 00:07:09,720 --> 00:07:13,480 Speaker 1: of concerns that information shared with the chatbot would not 110 00:07:13,680 --> 00:07:18,440 Speaker 1: be secure and thus violate citizen privacy laws. And we've 111 00:07:18,440 --> 00:07:22,760 Speaker 1: seen with chat GPT in particular that the history of 112 00:07:22,840 --> 00:07:27,080 Speaker 1: chat ended up becoming open right. People were suddenly able 113 00:07:27,120 --> 00:07:30,000 Speaker 1: to see what other people had been talking about with 114 00:07:30,120 --> 00:07:34,640 Speaker 1: chat GPT. So I think it's an understandable concern along 115 00:07:34,640 --> 00:07:38,800 Speaker 1: with privacy, but also it's good to see governments having 116 00:07:38,880 --> 00:07:43,760 Speaker 1: these discussions to really seriously talk about how to think 117 00:07:43,800 --> 00:07:46,400 Speaker 1: about AI in order to make the best use of 118 00:07:46,440 --> 00:07:51,920 Speaker 1: it and not have it create problems. Italy's approach has 119 00:07:51,920 --> 00:07:54,920 Speaker 1: motivated the EU in general. Lawmakers there as well as 120 00:07:54,920 --> 00:07:57,280 Speaker 1: some of the United States, have indicated that they are 121 00:07:57,320 --> 00:08:02,440 Speaker 1: now working on a code of conduct regarding AI and 122 00:08:02,840 --> 00:08:06,480 Speaker 1: AI companies. Representatives from the EU and the United States 123 00:08:06,520 --> 00:08:10,080 Speaker 1: met at a Trade and Technology Council meeting and started 124 00:08:10,120 --> 00:08:14,400 Speaker 1: to talk over this code of conduct, which would be voluntary. 125 00:08:15,280 --> 00:08:18,960 Speaker 1: On my mind, a voluntary code of conduct is a 126 00:08:18,960 --> 00:08:22,080 Speaker 1: little bit on the weak sauce side. I get that 127 00:08:22,760 --> 00:08:25,720 Speaker 1: optics can be bad if an AI company refuses to 128 00:08:25,760 --> 00:08:29,000 Speaker 1: adopt the code of conduct, and that might mean that 129 00:08:29,240 --> 00:08:32,120 Speaker 1: the company would find it difficult to do business with 130 00:08:32,200 --> 00:08:36,120 Speaker 1: customers if they refused to sign on to this code 131 00:08:36,160 --> 00:08:39,960 Speaker 1: of conduct. So there might be some social pressures and 132 00:08:40,080 --> 00:08:43,840 Speaker 1: business pressures to do it, but it's voluntary. Considering the 133 00:08:43,880 --> 00:08:47,880 Speaker 1: potential risks associated with AI now, I am not going 134 00:08:47,920 --> 00:08:51,199 Speaker 1: so far as to claim there's an existential threat level 135 00:08:51,280 --> 00:08:54,839 Speaker 1: risk out there, but there is risk, and that's bad enough. 136 00:08:54,880 --> 00:08:57,240 Speaker 1: But considering that risk, I think we might need more 137 00:08:57,280 --> 00:09:00,719 Speaker 1: than just a voluntary code of conduct in order to 138 00:09:00,800 --> 00:09:04,120 Speaker 1: keep things in line. Also, it'll be interesting to see 139 00:09:04,440 --> 00:09:08,600 Speaker 1: what role various AI companies and open Ai in particular, 140 00:09:08,960 --> 00:09:13,840 Speaker 1: will take in helping draft this code of conduct. You know, 141 00:09:13,880 --> 00:09:17,680 Speaker 1: Sam Altman, the CEO of open Ai, has tried to 142 00:09:17,679 --> 00:09:22,040 Speaker 1: get in front of this stuff and It makes me 143 00:09:22,120 --> 00:09:25,440 Speaker 1: worry because if the people who are the subject of 144 00:09:25,440 --> 00:09:29,280 Speaker 1: a code of conduct are actually allowed to write or 145 00:09:29,280 --> 00:09:32,760 Speaker 1: at least influence the writing of that code of conduct, 146 00:09:33,160 --> 00:09:35,319 Speaker 1: you can end up with rules that don't actually guard 147 00:09:35,320 --> 00:09:38,920 Speaker 1: against anything at all, and then it's just optics and 148 00:09:38,960 --> 00:09:42,920 Speaker 1: that's useless. And for a horrifying story of how AI 149 00:09:43,080 --> 00:09:47,360 Speaker 1: can be harmful, Chloe Shang of Motherboard has an article 150 00:09:47,400 --> 00:09:53,400 Speaker 1: titled Eating disorder Helpline disables chatbot for harmful responses after 151 00:09:53,559 --> 00:09:58,880 Speaker 1: firing human staff. Yeah, the headline, that's a lot. So 152 00:09:58,920 --> 00:10:02,000 Speaker 1: the story here is is that the National Eating Disorder 153 00:10:02,080 --> 00:10:08,000 Speaker 1: Association or NDA made a plan to replace human operators 154 00:10:08,360 --> 00:10:13,040 Speaker 1: of a mental health hotline with a chatbot called Tessup. 155 00:10:13,360 --> 00:10:15,840 Speaker 1: So I guess the idea was that the chatbot would 156 00:10:15,880 --> 00:10:20,040 Speaker 1: be more efficient and cheaper than keeping human beings who 157 00:10:20,080 --> 00:10:24,720 Speaker 1: have expertise and experience and you know, empathy on the payroll. 158 00:10:25,120 --> 00:10:27,840 Speaker 1: But you know, this is not a helpline you would 159 00:10:27,880 --> 00:10:30,640 Speaker 1: call if your lawnmower stopped working, Like, I can see 160 00:10:30,640 --> 00:10:34,400 Speaker 1: a chatbot being used for something rather mundane like this. 161 00:10:34,400 --> 00:10:37,360 Speaker 1: This is a helpline design for people who are dealing 162 00:10:38,000 --> 00:10:44,000 Speaker 1: with eating disorders. Union representatives have accused ANYDA of using 163 00:10:44,120 --> 00:10:47,760 Speaker 1: union busting tactics and warned that relying on AI could 164 00:10:47,840 --> 00:10:51,120 Speaker 1: lead to terrible situations. And earlier this week, a social 165 00:10:51,200 --> 00:10:53,640 Speaker 1: media post about how this AI chatbot led to a 166 00:10:53,720 --> 00:10:57,439 Speaker 1: terrible situation went viral. So, first up, there was an 167 00:10:57,440 --> 00:11:01,800 Speaker 1: activist named Sharon Maxwell. I had to test out Tessa, 168 00:11:02,400 --> 00:11:05,800 Speaker 1: and she said that quote every single thing Tessa suggested 169 00:11:05,840 --> 00:11:08,400 Speaker 1: were things that led to the development of my eating 170 00:11:08,440 --> 00:11:13,160 Speaker 1: disorder end quote. So by that, what I think Maxwell 171 00:11:13,200 --> 00:11:16,720 Speaker 1: is saying is that she had previously developed an eating disorder, 172 00:11:17,400 --> 00:11:21,280 Speaker 1: and the thoughts that went into her head that led 173 00:11:21,320 --> 00:11:24,280 Speaker 1: to her developing this eating disorder were the exact same 174 00:11:24,360 --> 00:11:28,440 Speaker 1: things that this chatbot was now suggesting as advice. So, 175 00:11:28,520 --> 00:11:32,840 Speaker 1: in other words, Tessa was giving Maxwell the advice of, Hey, 176 00:11:33,200 --> 00:11:36,440 Speaker 1: maybe you can lose a pound or two by doing this, 177 00:11:36,440 --> 00:11:39,200 Speaker 1: this and this, And when you're trying to make that 178 00:11:39,200 --> 00:11:42,520 Speaker 1: suggestion to someone who's dealing with the eating disorder, that 179 00:11:42,720 --> 00:11:46,560 Speaker 1: is a very dangerous thing. And then a psychologist named 180 00:11:46,600 --> 00:11:51,000 Speaker 1: Alexis Conason, and my apologies for the pronunciation of your 181 00:11:51,040 --> 00:11:53,520 Speaker 1: last name, I'm sure I'm getting it wrong. She conducted 182 00:11:53,559 --> 00:11:57,080 Speaker 1: her own test and found similar results. So what was 183 00:11:57,200 --> 00:12:03,520 Speaker 1: Anyda's response. Well, initially the organization accused Conison of fabricating 184 00:12:03,559 --> 00:12:07,400 Speaker 1: the whole thing, so she sent screenshots of her conversation 185 00:12:07,520 --> 00:12:10,320 Speaker 1: to an Eda, and then not too long after that, 186 00:12:10,880 --> 00:12:14,240 Speaker 1: Anyda took Tessa offline in order to address some quote 187 00:12:14,280 --> 00:12:19,120 Speaker 1: unquote bugs in the program. While Tesla has guardrails that 188 00:12:19,160 --> 00:12:22,160 Speaker 1: are meant to keep the chatbot from doing stuff like this, 189 00:12:23,120 --> 00:12:27,840 Speaker 1: we have seen again and again that AI can bypass 190 00:12:28,080 --> 00:12:30,880 Speaker 1: guardrails even if the person on the other end of 191 00:12:30,920 --> 00:12:35,120 Speaker 1: the conversation isn't trying to force things that way, and 192 00:12:35,240 --> 00:12:38,120 Speaker 1: the story really points out that for some jobs you 193 00:12:38,240 --> 00:12:41,200 Speaker 1: really probably should just depend upon human beings to do 194 00:12:41,240 --> 00:12:45,640 Speaker 1: the work. Okay, we're going to take a quick break. 195 00:12:45,679 --> 00:12:47,760 Speaker 1: When we come back, we've got some more news items 196 00:12:47,760 --> 00:13:02,040 Speaker 1: to cover. Okay, we're back. Meta says it's going to 197 00:13:02,120 --> 00:13:06,080 Speaker 1: remove posts containing news content for users in the state 198 00:13:06,080 --> 00:13:11,320 Speaker 1: of California if California passes a law that would require 199 00:13:11,400 --> 00:13:16,160 Speaker 1: platforms like Facebook and Google to pay publishers if work 200 00:13:16,200 --> 00:13:19,760 Speaker 1: from those publishers show up on those platforms. We've actually 201 00:13:19,760 --> 00:13:23,000 Speaker 1: seen this issue crop up around the world. Notably, it 202 00:13:23,080 --> 00:13:26,079 Speaker 1: happened in Australia a couple of years ago, when Australia 203 00:13:26,120 --> 00:13:30,199 Speaker 1: passed a similar law, Facebook went dark in Australia temporarily. 204 00:13:30,520 --> 00:13:34,400 Speaker 1: Eventually the law took hold and things have kind of 205 00:13:34,520 --> 00:13:37,720 Speaker 1: entered into an equilibrium. Whether or not the law actually 206 00:13:37,760 --> 00:13:42,280 Speaker 1: addressed the issue that was of concern is another matter 207 00:13:42,520 --> 00:13:46,120 Speaker 1: that actually is a good subject for an episode, honestly. 208 00:13:46,520 --> 00:13:50,240 Speaker 1: But the idea is that publishers want compensation. They're arguing 209 00:13:50,760 --> 00:13:55,520 Speaker 1: that platforms like Google and Facebook are siphoning traffic away 210 00:13:55,600 --> 00:13:59,880 Speaker 1: from the actual news websites. That's where they monetize that traffic, 211 00:14:00,440 --> 00:14:04,760 Speaker 1: and these platforms are benefiting from the work being done 212 00:14:04,800 --> 00:14:08,920 Speaker 1: by journalists, but they're not compensating the news outlets in 213 00:14:08,960 --> 00:14:13,760 Speaker 1: the process. Meta's Andy Stone, who's a name that I 214 00:14:13,800 --> 00:14:16,400 Speaker 1: see pop up pretty much whenever the company wants to 215 00:14:16,400 --> 00:14:20,400 Speaker 1: dismiss regulations that would work against it, said the bill, 216 00:14:20,520 --> 00:14:22,840 Speaker 1: if signed into law, would amount to nothing more than 217 00:14:22,840 --> 00:14:26,480 Speaker 1: a slush fund that would benefit large media companies but 218 00:14:26,640 --> 00:14:30,840 Speaker 1: not help smaller California based publishers, which, depending on how 219 00:14:30,880 --> 00:14:33,800 Speaker 1: this bill is framed, that actually might be true, because 220 00:14:33,800 --> 00:14:37,880 Speaker 1: I've heard similar criticisms about the Australian law. Now I 221 00:14:37,920 --> 00:14:40,160 Speaker 1: don't know I haven't read the bill. If I had 222 00:14:40,240 --> 00:14:42,960 Speaker 1: read the bill, I probably wouldn't have a good grasp 223 00:14:43,000 --> 00:14:46,440 Speaker 1: on its limitations because law speak be scary you. It's 224 00:14:46,440 --> 00:14:50,160 Speaker 1: even more difficult to read than a really complex technical manual. 225 00:14:50,600 --> 00:14:54,600 Speaker 1: But yes, this is another battle we're starting to see unfold. 226 00:14:55,040 --> 00:14:59,640 Speaker 1: I don't see California backing down from this, and it 227 00:14:59,680 --> 00:15:02,800 Speaker 1: will be interesting to see where this goes. But honestly, 228 00:15:02,880 --> 00:15:05,520 Speaker 1: at the end of the day, I'm mostly concerned that 229 00:15:05,600 --> 00:15:09,400 Speaker 1: the law actually does what the law aims to do right. 230 00:15:10,160 --> 00:15:13,040 Speaker 1: It gets very frustrating when you hear about laws that 231 00:15:13,120 --> 00:15:17,200 Speaker 1: potentially could correct a situation, but because of how they 232 00:15:17,200 --> 00:15:20,600 Speaker 1: are written and how they're enacted, they failed to do 233 00:15:21,000 --> 00:15:26,520 Speaker 1: what the purpose at least claimed to be. Metta also 234 00:15:26,840 --> 00:15:29,880 Speaker 1: faces a fine levied by a court in Russia this week. 235 00:15:30,240 --> 00:15:33,080 Speaker 1: The charge is that the company failed to remove prohibited 236 00:15:33,160 --> 00:15:38,560 Speaker 1: content from WhatsApp, specifically about a drug called Lyrica, and 237 00:15:38,600 --> 00:15:44,160 Speaker 1: so the fine is three million rubles, which equates to 238 00:15:44,240 --> 00:15:47,720 Speaker 1: about thirty seven thousand dollars, and a Russian court also 239 00:15:47,840 --> 00:15:52,560 Speaker 1: find the Wikimedia Foundation a similar amount of money, also 240 00:15:52,680 --> 00:15:56,560 Speaker 1: three million rubles, but they said that Wikimedia Foundation failed 241 00:15:56,560 --> 00:16:00,720 Speaker 1: to remove quote unquote false information about Russia's war wind Ukraine. 242 00:16:01,560 --> 00:16:04,960 Speaker 1: Something tells me that neither Meta nor Wikimedia Foundation will 243 00:16:05,000 --> 00:16:09,840 Speaker 1: consider these moves particularly intimidating. Meta certainly not thirty seven 244 00:16:09,880 --> 00:16:12,960 Speaker 1: thousand dollars is like, I don't even think they would 245 00:16:13,000 --> 00:16:16,440 Speaker 1: notice if that money went away, So I don't think 246 00:16:16,440 --> 00:16:20,080 Speaker 1: that this is really a big move against the companies. 247 00:16:20,600 --> 00:16:23,440 Speaker 1: Amazon also has a couple of bills to settle this week. 248 00:16:23,640 --> 00:16:25,640 Speaker 1: First up, the company agreed to pay a thirty eight 249 00:16:25,680 --> 00:16:28,680 Speaker 1: million dollars settlement in relation to a lawsuit that accused 250 00:16:28,720 --> 00:16:33,000 Speaker 1: the company of having illegally collected and storing information relating 251 00:16:33,040 --> 00:16:39,200 Speaker 1: to children through the Digital Personal Assistant, whose name starts 252 00:16:39,200 --> 00:16:41,520 Speaker 1: with A, ends with A, and has lex in the middle. 253 00:16:42,200 --> 00:16:45,560 Speaker 1: The Federal Trade Commission in the United States brought the 254 00:16:45,640 --> 00:16:49,080 Speaker 1: case against Amazon, so this settlement requires that the company 255 00:16:49,400 --> 00:16:53,640 Speaker 1: changes its data collection, storage, and deletion practices on top 256 00:16:53,680 --> 00:16:56,360 Speaker 1: of paying the fine. The other bill Amazon has to 257 00:16:56,360 --> 00:16:59,920 Speaker 1: pay is five point eight million dollars. This is another settlement, 258 00:17:00,160 --> 00:17:03,800 Speaker 1: also with the FTC. This one is with regard to 259 00:17:03,880 --> 00:17:08,200 Speaker 1: Amazon's ring products. Those are the security systems and doorbell 260 00:17:08,560 --> 00:17:12,200 Speaker 1: camera systems. The FTC accused Amazon of having a system 261 00:17:12,240 --> 00:17:17,120 Speaker 1: that allowed employees and contractors to access video feeds from 262 00:17:17,200 --> 00:17:21,760 Speaker 1: customer cameras without any real safeguards to prevent that from happening. 263 00:17:22,160 --> 00:17:24,959 Speaker 1: As you might imagine that as a huge privacy and 264 00:17:25,000 --> 00:17:29,000 Speaker 1: security violation. Amazon has agreed to create new processes with 265 00:17:29,119 --> 00:17:31,159 Speaker 1: regular checkups to make sure that the company has a 266 00:17:31,160 --> 00:17:35,120 Speaker 1: tighter data security strategy, while also simultaneously saying we never 267 00:17:35,160 --> 00:17:38,880 Speaker 1: broke the law, because that's what you can do when 268 00:17:38,920 --> 00:17:42,119 Speaker 1: you make settlements, all right. One trend happening with some 269 00:17:42,280 --> 00:17:45,000 Speaker 1: tech platforms is to make changes to how these platforms 270 00:17:45,000 --> 00:17:48,920 Speaker 1: give access to an API, which is an application programming interface. 271 00:17:49,400 --> 00:17:53,800 Speaker 1: That's what lets app developers tap into larger platforms in 272 00:17:53,920 --> 00:17:57,240 Speaker 1: order to do whatever it is. So a developer might 273 00:17:57,320 --> 00:18:00,240 Speaker 1: create an app that ties into a larger platform like 274 00:18:00,280 --> 00:18:02,760 Speaker 1: Twitter or as we'll talk about in a second Reddit, 275 00:18:03,440 --> 00:18:08,119 Speaker 1: and these apps then reference they send requests to the 276 00:18:08,640 --> 00:18:13,320 Speaker 1: underlying platform, and that populates the app. So Twitter as 277 00:18:13,480 --> 00:18:17,320 Speaker 1: a price tag associated with this. For every fifty million 278 00:18:17,400 --> 00:18:21,200 Speaker 1: tweets requested, a developer has to cough up forty two 279 00:18:21,280 --> 00:18:24,800 Speaker 1: thousand dollars. And yes, fifty million tweets, that's a lot. 280 00:18:25,080 --> 00:18:27,760 Speaker 1: But if your app is super popular and a lot 281 00:18:27,760 --> 00:18:30,119 Speaker 1: of people are using it all day. You're going to 282 00:18:30,200 --> 00:18:33,240 Speaker 1: hit that fifty million mark over and over again. Reddit 283 00:18:33,400 --> 00:18:36,480 Speaker 1: is now doing something similar. They've changed their API, and 284 00:18:36,640 --> 00:18:41,119 Speaker 1: Christian Sellig, the developer behind the popular Reddit app Apollo, 285 00:18:41,680 --> 00:18:43,800 Speaker 1: reports that he might have to shut the app down 286 00:18:43,960 --> 00:18:46,919 Speaker 1: entirely because of Reddit's new policy, which is to charge 287 00:18:46,960 --> 00:18:51,439 Speaker 1: twelve thousand dollars per fifty million requests. Selleg revealed that 288 00:18:51,480 --> 00:18:56,199 Speaker 1: Apollo generates around seven billion requests every month, and that 289 00:18:56,240 --> 00:18:58,800 Speaker 1: means it would cost around twenty million dollars a year 290 00:18:58,840 --> 00:19:02,400 Speaker 1: to operate. The app, understandably is not in a position 291 00:19:02,600 --> 00:19:06,200 Speaker 1: to pay that much. Generally speaking, the app developer community 292 00:19:06,280 --> 00:19:08,360 Speaker 1: is not too keen on this approach, as it punishes 293 00:19:08,400 --> 00:19:12,040 Speaker 1: you for being successful. Finally, if you have a PC 294 00:19:12,200 --> 00:19:14,760 Speaker 1: rig with a motherboard from Gigabyte, you should know that 295 00:19:14,800 --> 00:19:18,000 Speaker 1: security researchers at Eclipsium discovered the company had created a 296 00:19:18,040 --> 00:19:22,000 Speaker 1: backdoor system to deliver firmware updates to the motherboard, and 297 00:19:22,480 --> 00:19:25,520 Speaker 1: that system lacks proper security, which means a hacker could 298 00:19:25,520 --> 00:19:29,920 Speaker 1: potentially hijack that delivery system and use it to send 299 00:19:30,520 --> 00:19:34,720 Speaker 1: executable code straight to target computers. Do not pass goes, 300 00:19:34,760 --> 00:19:37,479 Speaker 1: do not Colick two hundred dollars. If you're curious, if 301 00:19:37,520 --> 00:19:40,399 Speaker 1: your device has a Gigabyte motherboard, you can go to 302 00:19:41,080 --> 00:19:43,680 Speaker 1: the start menu and Windows and look at system information. 303 00:19:44,800 --> 00:19:48,000 Speaker 1: More than two hundred and fifty motherboards are affected by this, 304 00:19:48,760 --> 00:19:52,439 Speaker 1: so that's a big ouch. Supposedly, Gigabyte is working on 305 00:19:52,520 --> 00:19:56,320 Speaker 1: this and intends to create a solution, but as of 306 00:19:56,400 --> 00:19:59,240 Speaker 1: right now, I don't know of any solution that has 307 00:19:59,359 --> 00:20:03,320 Speaker 1: actually out there, so be careful out there. All right, 308 00:20:03,400 --> 00:20:05,960 Speaker 1: that's it for the tech news for today, June first, 309 00:20:06,000 --> 00:20:09,560 Speaker 1: twenty twenty three. I hope you're all well, and I'll 310 00:20:09,600 --> 00:20:18,760 Speaker 1: talk to you again really soon. Tech Stuff is an 311 00:20:18,840 --> 00:20:24,359 Speaker 1: iHeartRadio production. For more podcasts from iHeartRadio, visit the iHeartRadio app, 312 00:20:24,480 --> 00:20:31,560 Speaker 1: Apple Podcasts, or wherever you listen to your favorite shows.