WEBVTT - Tech News: Exploding Pagers and Nuclear-Powered AI 0:00:04.480 --> 0:00:12.319 Welcome to tech Stuff, a production from iHeartRadio. Hey there, 0:00:12.440 --> 0:00:16.160 and welcome to tech Stuff. I'm your host job and Strickland, 0:00:16.160 --> 0:00:20.000 I'm an executive producer with iHeart Podcasts. And how the 0:00:20.160 --> 0:00:22.560 tech are you? It's time for the tech news for 0:00:22.640 --> 0:00:27.280 the week ending September twenty, twenty twenty four. The US 0:00:27.360 --> 0:00:31.760 Federal Trade Commission or FTC, issued a release this week 0:00:31.800 --> 0:00:35.479 titled well, actually it's a really long title. Essentially, this 0:00:35.600 --> 0:00:40.320 release says social media and streaming companies are essentially conducting 0:00:40.400 --> 0:00:45.360 surveillance on both their users and on non users while 0:00:45.400 --> 0:00:51.040 also maintaining inadequate data security and privacy controls, Which I mean, 0:00:51.680 --> 0:00:54.320 if you're someone who has even paid just a little 0:00:54.360 --> 0:00:57.760 attention to these companies over the last several years, you 0:00:57.880 --> 0:01:01.040 might respond to this news as oh, yeah, yeah yeah. Also, 0:01:01.160 --> 0:01:03.840 by the way, water is wet. At least that was 0:01:03.920 --> 0:01:07.679 my initial reaction, because to me, the revelation that these 0:01:07.840 --> 0:01:12.120 companies are one collecting massive amounts of data about people 0:01:12.400 --> 0:01:16.680 and two they're pretty crappy about keeping that stuff safe, 0:01:16.880 --> 0:01:21.040 that's not exactly news. But then the FTC was conducting 0:01:21.080 --> 0:01:24.960 an investigation to determine exactly the scope of data collection 0:01:25.520 --> 0:01:28.840 as well as to what extent, if any, these companies 0:01:28.880 --> 0:01:32.600 are taking to you know, protect private information. These are 0:01:32.640 --> 0:01:34.480 the sorts of things that you got to do if 0:01:34.560 --> 0:01:37.800 maybe further down the line, you decide to you know, 0:01:37.880 --> 0:01:42.600 pressure companies into making changes or pressuring the government into 0:01:42.640 --> 0:01:46.479 passing laws that will require companies to make changes. Now, 0:01:46.520 --> 0:01:50.120 perhaps by by finding the ever loven socks off these 0:01:50.200 --> 0:01:54.200 companies a guy I can dream anyway. The release detailed 0:01:54.280 --> 0:01:58.640 how these companies collect information not just about their own users, 0:01:58.800 --> 0:02:02.160 I mean that is evident, but also on people who 0:02:02.200 --> 0:02:05.120 aren't users at all. And let's just say, how, you know, 0:02:05.280 --> 0:02:07.000 how could that happen? How could someone who's not a 0:02:07.080 --> 0:02:09.920 user have their data collected by these companies. Well, let's 0:02:10.240 --> 0:02:14.000 give one simple example. Let's say you gut Uncle Joey, 0:02:14.280 --> 0:02:17.519 and Uncle Joey's not on Facebook, but you do post 0:02:17.520 --> 0:02:20.440 about Uncle Joey a lot on Facebook, and you know, 0:02:20.639 --> 0:02:24.480 Facebook now knows stuff about Uncle Joey. That's one way. 0:02:24.960 --> 0:02:28.359 Another is that companies might get access to people's information 0:02:28.480 --> 0:02:31.680 by dealing with data brokers. Data brokers are companies that 0:02:31.880 --> 0:02:34.840 buy and sell personal information across the Internet, and a 0:02:34.919 --> 0:02:38.400 data broker might have quite the dossier on you or 0:02:38.440 --> 0:02:42.120 on Uncle Joey, and that information has gathered from multiple 0:02:42.160 --> 0:02:46.120 sources and organized into handy dandy packets for any company 0:02:46.120 --> 0:02:49.239 that really wants to know you know, like what kind 0:02:49.400 --> 0:02:54.080 of breakfast cereal you like, or what kind of car 0:02:54.520 --> 0:02:58.080 you drive, or which political party you support, or whether 0:02:58.200 --> 0:03:01.240 or not you like sports that kind of thing. The 0:03:01.400 --> 0:03:05.480 FTC found most of these companies lack sufficient data security 0:03:05.520 --> 0:03:08.840 and retention policies, meaning that some of these companies retain 0:03:08.919 --> 0:03:13.000 personal information indefinitely. So back in the corporate world, there 0:03:13.000 --> 0:03:17.520 are often rules that mandate that companies destroy older documents 0:03:17.560 --> 0:03:20.520 as time passes. I worked for a consulting company once 0:03:20.600 --> 0:03:23.040 upon a time, and they had strict rules on how 0:03:23.040 --> 0:03:26.359 long we could retain files on our customers, for example, 0:03:26.600 --> 0:03:29.840 and once that time was up, we were obligated to 0:03:30.000 --> 0:03:34.080 delete electronic files and shred hard copies. This was for 0:03:34.200 --> 0:03:37.680 the protection both for the customer and for the consulting 0:03:37.720 --> 0:03:42.600 firm itself. But these social network companies and social streaming sites, 0:03:42.880 --> 0:03:45.680 they often don't have that kind of a policy, and 0:03:45.760 --> 0:03:48.840 so the information to have on people can stay locked 0:03:48.880 --> 0:03:52.720 away and grow over time. The report or just Congress 0:03:52.720 --> 0:03:55.560 to pass laws that will give citizens more data security, 0:03:56.040 --> 0:03:59.560 maybe even give citizens a little bit of agency when 0:03:59.560 --> 0:04:02.440 it comes to how their own personal data can be 0:04:02.440 --> 0:04:06.080 collected and used. Wouldn't that be nice? And further that 0:04:06.160 --> 0:04:10.880 Congress should prioritize protecting younger Internet users in particular, who 0:04:10.880 --> 0:04:15.480 often end up being exploited due to massive loopholes and systems. 0:04:15.560 --> 0:04:19.320 We talked about this recently about how companies like Google 0:04:19.400 --> 0:04:25.880 through YouTube were allegedly working with advertisers to target teenagers, 0:04:25.960 --> 0:04:29.880 even though that's expressly against the rules, because what they 0:04:29.880 --> 0:04:33.440 could do is say, well, let's just target unknown accounts 0:04:33.839 --> 0:04:37.200 where we don't have an age associated with the account holder, 0:04:37.440 --> 0:04:43.719 but the account has behaviors that are typically associated with, say, teenagers, right, 0:04:44.040 --> 0:04:47.840 So you're not targeting teenagers. You're targeting people who behave 0:04:48.040 --> 0:04:50.520 like teenagers and you don't know how old they are, 0:04:50.600 --> 0:04:54.120 So how could you be blamed for targeting teenagers? That's 0:04:54.160 --> 0:04:57.800 the allegation that's going around regarding YouTube. So loopholes like 0:04:57.839 --> 0:05:02.599 that provide huge op oportunities for these companies to exploit 0:05:02.760 --> 0:05:06.240 people that otherwise are meant to be protected from that 0:05:06.360 --> 0:05:09.480 sort of thing. So the FTC actually voted with unanimous 0:05:09.480 --> 0:05:12.080 support to issue this report I think that's pretty refreshing 0:05:12.360 --> 0:05:15.800 for an agency to say, yeah, we're all aligned on this, 0:05:15.960 --> 0:05:18.880 we all think that this is important, and you don't 0:05:18.920 --> 0:05:22.440 get that division you typically get with these where out 0:05:22.440 --> 0:05:28.000 of just principle members will vote against a measure because 0:05:28.560 --> 0:05:32.000 unity is something to be avoided at all costs. Enough 0:05:32.000 --> 0:05:35.800 with that commentary. One huge story that unfolded over this 0:05:35.880 --> 0:05:39.760 past week involved pagers or what we used to call 0:05:39.920 --> 0:05:42.159 beepers when I was a kid, and I'm sure you've 0:05:42.160 --> 0:05:46.560 heard the disturbing and grizzly details. So this week pagers 0:05:46.560 --> 0:05:50.760 in Lebanon and Syria began to explode and they killed 0:05:50.880 --> 0:05:53.760 or at least a dozen people and injured thousands in 0:05:53.800 --> 0:05:57.400 the process. The following day, another series of explosions rocked 0:05:57.400 --> 0:06:01.040 the area, this time from walkie talkie hands sets blowing up. 0:06:01.240 --> 0:06:04.440 The operation is believed to be a semi targeted attack 0:06:04.600 --> 0:06:09.120 on Hesblon that was orchestrated by Israeli spies. Now I 0:06:09.240 --> 0:06:12.800 say semi targeted because it's pretty darn hard to guarantee 0:06:12.800 --> 0:06:14.880 a device like that it's going to be on the 0:06:14.920 --> 0:06:18.800 actual target that you have in mind, or it's really 0:06:18.839 --> 0:06:22.760 impossible to guarantee that that target isn't also near innocent 0:06:22.839 --> 0:06:26.880 civilians when the device actually does explode. In fact, reportedly 0:06:27.080 --> 0:06:30.600 some of the casualties have been children, which is obviously 0:06:30.640 --> 0:06:33.880 it's truly heartbreaking. It's a horrible thing. We do have 0:06:33.920 --> 0:06:37.640 some more details courtesy of the New York Times. The article, 0:06:37.680 --> 0:06:41.440 which is titled how Israel built a modern day trojan 0:06:41.520 --> 0:06:46.440 Horse exploding Pagers gives more details as to what actually happened. 0:06:46.800 --> 0:06:50.280 So the pagers contained a small amount of explosives, and 0:06:50.320 --> 0:06:54.239 that explosive would activate upon receiving a signal. That signal 0:06:54.320 --> 0:06:55.960 was a message that was sent out at three point 0:06:55.960 --> 0:06:59.400 thirty pm in local time. The New York Times article 0:06:59.440 --> 0:07:02.800 reports that Israel has been planning this operation for a 0:07:02.839 --> 0:07:06.839 long time. HESBLA, which obviously has been in a violent 0:07:06.960 --> 0:07:11.480 conflict with Israel for decades now, has more recently attempted 0:07:11.520 --> 0:07:16.520 to migrate members to lower tech communications solutions. That's in 0:07:16.520 --> 0:07:20.440 an effort to remain undetected by Israeli agents. The argument 0:07:20.560 --> 0:07:24.840 was that smartphones and things like that they were helping 0:07:25.200 --> 0:07:30.880 Israeli military and spies target people in HESBLA, So in 0:07:30.960 --> 0:07:34.400 order to avoid that, switch to lower tech stuff that 0:07:34.440 --> 0:07:37.640 doesn't have the same capacity for being tracked. That was 0:07:37.760 --> 0:07:42.040 the concept. Israel, however, knew about this and apparently took 0:07:42.040 --> 0:07:46.280 this opportunity to build such devices with explosives directly incorporated 0:07:46.320 --> 0:07:51.200 into them, and then essentially fed these products to Hesbela 0:07:51.320 --> 0:07:55.280 agents through third parties. So Israel just made sure that 0:07:55.320 --> 0:08:01.040 these other companies had these explosive devices, and then you know, 0:08:01.120 --> 0:08:05.200 when when Hesbela agents were looking for these low tech 0:08:05.240 --> 0:08:07.840 communications devices, the ones they got were the ones that 0:08:07.880 --> 0:08:11.280 were built by Israel. The attack sounded like something that 0:08:11.320 --> 0:08:13.760 would be in a far fetched action movie, you know, 0:08:13.800 --> 0:08:15.960 something you might see in like one of the Kingsman 0:08:16.120 --> 0:08:20.200 films or something. And I cannot imagine how horrifying and 0:08:20.400 --> 0:08:23.800 terrifying it must have been to have been present when 0:08:23.840 --> 0:08:28.080 those explosions began to happen. It's a ready sobering, i 0:08:28.080 --> 0:08:31.680 mean beyond sobering thing to think about, and kind of 0:08:31.800 --> 0:08:36.640 terrifying really, just that this was a coordinated effort that 0:08:36.840 --> 0:08:39.880 was extremely successful in hurting a lot of people. Now, 0:08:39.960 --> 0:08:43.280 whether those were the quote unquote right people, I don't know. 0:08:43.600 --> 0:08:47.600 I'm generally against people hurting each other at all, so 0:08:48.320 --> 0:08:51.880 I'm not big fan on Hesbela carrying out acts of 0:08:51.920 --> 0:08:55.800 violence against people in Israel. That's horrible. Not really keen 0:08:55.880 --> 0:08:59.840 on Israeli military and spies carrying out acts of violence 0:08:59.840 --> 0:09:03.199 on people in Lebanon in Syria. That's horrible. It's all terrible. 0:09:03.320 --> 0:09:06.319 That's really what I get down to it. Over in Europe, 0:09:06.640 --> 0:09:09.840 Reuter's reports that major tech companies are trying to convince 0:09:09.880 --> 0:09:13.439 EU leaders to not go so heavy handed with regard 0:09:13.520 --> 0:09:18.079 to upcoming laws and regulations relating to artificial intelligence. Now, 0:09:18.120 --> 0:09:22.319 you might recall that open aiy's Sam Altman started meeting 0:09:22.400 --> 0:09:26.400 with various political leaders like a couple of years ago. Now, 0:09:26.640 --> 0:09:31.480 those were in early discussions about proposed AI regulations, and 0:09:31.600 --> 0:09:35.040 Altman reportedly was calling four rules. He was saying, yes, 0:09:35.080 --> 0:09:38.719 we should have regulations for AI. However, critics accused him 0:09:38.760 --> 0:09:41.800 of trying to lay a foundation that would really benefit 0:09:42.080 --> 0:09:48.000 open ai while simultaneously suppressing potential competition from smaller companies, 0:09:48.360 --> 0:09:51.440 and all the while, the actual rules that were created 0:09:51.440 --> 0:09:55.280 would do very little to protect the citizens of the EU. 0:09:56.080 --> 0:09:59.480 Even if the tech companies failed to persuade legislators to 0:09:59.559 --> 0:10:04.040 use a lighter touch when it comes to creating AI policy, 0:10:04.440 --> 0:10:07.200 they don't have too much to be worried about, at 0:10:07.280 --> 0:10:10.240 least in the short term, because according to Reuter's the 0:10:10.360 --> 0:10:13.880 rules that are being discussed now, once they are adopted, 0:10:13.920 --> 0:10:17.760 they aren't actually legally binding. They would be, as Captain 0:10:17.800 --> 0:10:23.400 Barbosa might say, more like guidelines also a happy belated 0:10:23.520 --> 0:10:25.960 talk like a pirate day. I'm a day late on 0:10:25.960 --> 0:10:29.240 that one. Anyway, if a company is found to ignore 0:10:29.440 --> 0:10:33.800 or violate guidelines, particularly multiple times, then it could potentially 0:10:33.800 --> 0:10:37.040 face a more serious challenge from regulators in the future, 0:10:37.160 --> 0:10:40.320 which ultimately could end up with like fines and that 0:10:40.400 --> 0:10:43.560 kind of stuff. The issues at play here include everything 0:10:43.640 --> 0:10:47.200 from how companies will gather data to train their AI 0:10:47.400 --> 0:10:50.679 like data scraping. In other words, there are concerns about copyright, 0:10:50.840 --> 0:10:55.080 like can AI companies indiscriminately train models on information that 0:10:55.440 --> 0:10:59.160 otherwise has the protection from such use. Transparency is a 0:10:59.160 --> 0:11:01.760 really big part of it, because right now AI largely 0:11:01.800 --> 0:11:05.959 follows the black box model of transparency, meaning there is 0:11:06.000 --> 0:11:09.480 no transparency. Stuff goes in, other stuff comes out. We're 0:11:09.480 --> 0:11:12.160 not allowed to see what process happens in between those 0:11:12.160 --> 0:11:15.480 two things, that kind of thing. Reuters reveals that many 0:11:15.520 --> 0:11:19.320 major companies, including Google and Open Ai, are lobbying to 0:11:19.360 --> 0:11:22.640 be included in working groups that are meant to shape policies, 0:11:22.679 --> 0:11:24.920 which again gets back to that issue I talked about 0:11:24.920 --> 0:11:27.280 a moment ago. When the folks who are meant to 0:11:27.320 --> 0:11:30.720 be regulated are allowed to shape the regulations, what do 0:11:30.800 --> 0:11:32.800 you think is going to happen. Who do you think 0:11:32.840 --> 0:11:36.120 that's going to ultimately benefit? Here's a hint, it's not 0:11:36.240 --> 0:11:39.120 you or me. Okay, we've got more tech news to 0:11:39.160 --> 0:11:50.160 get through, but first let's take a quick break. The 0:11:50.280 --> 0:11:55.080 United Nations has issued a report that recommends the organization 0:11:55.440 --> 0:11:58.319 of a committee that would be in charge of governing 0:11:58.360 --> 0:12:02.280 and monitoring the development and ployment of artificial intelligence around 0:12:02.320 --> 0:12:05.640 the world. The report suggests that this committee would have 0:12:05.720 --> 0:12:09.960 a similar structure and status that the Intergovernmental Panel on 0:12:10.080 --> 0:12:14.960 Climate Change possesses. The recommendations aren't all about putting AI 0:12:15.040 --> 0:12:20.560 in the spotlight as a potentially dangerous, perhaps catastrophically dangerous technology. 0:12:20.920 --> 0:12:24.800 It's also to kind of decide how to leverage AI 0:12:24.960 --> 0:12:29.040 in ways that provide the most benefit while potentially creating 0:12:29.400 --> 0:12:33.200 the least potential for harm. How to ensure that poorer 0:12:33.280 --> 0:12:36.880 countries are able to realize the benefits from AI while 0:12:36.880 --> 0:12:38.559 also having a seat at the table when it comes 0:12:38.600 --> 0:12:40.840 to governance and thus not be left behind, not to 0:12:40.880 --> 0:12:45.000 create an AI gap. In other words, as we have seen, 0:12:45.679 --> 0:12:49.920 artificial intelligence is really complicated, right, It's not just good 0:12:50.160 --> 0:12:54.000 or bad. There's a lot of complex stuff going on here, 0:12:54.520 --> 0:12:58.640 and it's far too easy to fall into simplifying things 0:12:58.720 --> 0:13:02.840 beyond reas I've done that personally. I mean in just 0:13:02.880 --> 0:13:05.360 in an effort to try and communicate stuff, I have 0:13:05.480 --> 0:13:10.880 oversimplified the problem. I recommend Will Knight's article over on Wired. 0:13:11.120 --> 0:13:14.400 It's titled the United Nations wants to treat AI with 0:13:14.480 --> 0:13:18.079 the same urgency as climate change. Knight does a great 0:13:18.160 --> 0:13:22.079 job explaining not only the recent report that the United 0:13:22.160 --> 0:13:25.720 Nations issued, but also the existing regulations that are already 0:13:25.760 --> 0:13:30.560 in place that could potentially shape this intergovernmental body as 0:13:30.600 --> 0:13:34.719 it takes form. So check that out. Emily Bernbaum and 0:13:34.880 --> 0:13:39.559 Oma Sadek of Bloomberg have a piece that's titled Microsoft 0:13:39.679 --> 0:13:43.600 executive Warrens of Election Meddling in Final forty eight hours. 0:13:44.000 --> 0:13:48.000 So the referenced executive is Brad Smith, who told the 0:13:48.080 --> 0:13:51.840 Senate Intelligence Committee here in the United States that foreign 0:13:51.960 --> 0:13:56.520 backed campaigns aimed to interfere with US elections are likely 0:13:56.559 --> 0:14:00.959 going to peak two days before the election itself. Examples 0:14:01.000 --> 0:14:04.360 that have happened in other countries in which deep fakes 0:14:04.360 --> 0:14:07.920 of various candidates in those elections made the rounds just 0:14:08.040 --> 0:14:11.960 before their election day. We've already seen some examples of 0:14:12.000 --> 0:14:14.920 AI generated material relating to the election here in the 0:14:15.040 --> 0:14:18.280 United States, though for the most part it hasn't actually 0:14:18.320 --> 0:14:22.080 made a huge impact, although one example did prompt Taylor 0:14:22.160 --> 0:14:25.360 Swift to make a statement about voting and also the 0:14:25.400 --> 0:14:29.720 candidate she personally supports, partly because an AI generated image 0:14:29.760 --> 0:14:32.840 of her appearing to support Donald Trump got some buzz, 0:14:33.200 --> 0:14:37.320 especially after Trump himself boosted that signal, and Swift's response 0:14:37.400 --> 0:14:40.040 has inspired a few hundred thousand citizens here in the 0:14:40.200 --> 0:14:42.360 US to register to vote. So that's actually a really 0:14:42.400 --> 0:14:45.840 good thing, you know. It's participation in democracy is actually 0:14:45.840 --> 0:14:49.760 absolutely vital for democracy success. But yeah, we're likely to 0:14:49.760 --> 0:14:52.840 see a lot more of that leading up to election day, 0:14:52.840 --> 0:14:56.160 according to Smith, and that sounds like it tracks to me. 0:14:56.520 --> 0:14:59.480 We've already seen multiple stories about how countries like Iran 0:14:59.680 --> 0:15:02.640 and Russia have attempted to shape the election to varying 0:15:02.680 --> 0:15:06.440 degrees of success here, especially this past year, So for 0:15:06.520 --> 0:15:10.320 that to escalate seems like it's a pretty safe bet. 0:15:10.960 --> 0:15:14.200 Unsafe consequences, but a safe bet that it's gonna happen. 0:15:14.840 --> 0:15:18.480 Simon Sharwood of The Register has an article titled LinkedIn 0:15:18.680 --> 0:15:22.600 started harvesting people's posts for training AI without asking for 0:15:22.680 --> 0:15:26.120 opt in. And again, I think that probably doesn't come 0:15:26.120 --> 0:15:28.800 as a surprise to most people. I mean, lots of 0:15:28.840 --> 0:15:32.760 platforms have done this, where they started to scrape their 0:15:32.800 --> 0:15:36.800 own platforms for user information to train their AI models 0:15:37.160 --> 0:15:41.520 without notifying users, let alone asking them for their consent 0:15:41.800 --> 0:15:45.240 for this to happen. But while it might not be surprising, 0:15:45.280 --> 0:15:47.600 it's still very upsetting to a lot of people who 0:15:47.840 --> 0:15:51.400 likely did not anticipate that their online resume and their 0:15:51.480 --> 0:15:55.480 various posts promoting their work or their companies would ultimately 0:15:55.520 --> 0:15:59.280 serve as training fodder for AI models. Ashley Bellinger of 0:15:59.320 --> 0:16:02.080 Ours Teneca has a related piece to this. It is 0:16:02.160 --> 0:16:06.240 titled how to stop LinkedIn from training AI on your 0:16:06.320 --> 0:16:09.760 data Now. Bellinger points out right away that there's no 0:16:09.840 --> 0:16:12.600 way to opt out of any training that has already happened. 0:16:12.840 --> 0:16:15.880 You cannot say, oh, you know what, remove all my 0:16:16.000 --> 0:16:20.200 stuff from your training model's brain. That's no go. It's 0:16:20.280 --> 0:16:24.120 that cat is out of the AI bag already, so 0:16:24.240 --> 0:16:28.000 there's no way to protect your data that has already 0:16:28.040 --> 0:16:32.160 been used to train LinkedIn AI models. However, there are 0:16:32.240 --> 0:16:36.840 some somewhat limited ways to opt out of future training. 0:16:37.240 --> 0:16:40.800 LinkedIn will issue an updated user agreement related to this, 0:16:41.040 --> 0:16:43.360 and at that point users will be able to opt 0:16:43.400 --> 0:16:46.200 out of future training sessions. To do that, you will 0:16:46.200 --> 0:16:48.640 need to go to your data privacy settings and look 0:16:48.680 --> 0:16:51.160 for a bit that's related to data collection for quote 0:16:51.200 --> 0:16:55.440 unquote generative AI improvement and make sure that that option 0:16:55.560 --> 0:16:59.600 is turned off. Now. If you are in Switzerland or 0:16:59.640 --> 0:17:03.080 if you in the EU, where the law by default 0:17:03.200 --> 0:17:07.359 requires LinkedIn to secure your consent before opting you into 0:17:07.480 --> 0:17:10.040 this data collection program in the first place, you don't 0:17:10.080 --> 0:17:12.879 have to worry about this. You will be able to 0:17:13.200 --> 0:17:16.600 actually respond no, don't do that when you are prompted. 0:17:16.840 --> 0:17:19.360 The rest of the world we don't get that treatment. 0:17:19.680 --> 0:17:22.080 But hey, I think that's a great example to bring 0:17:22.160 --> 0:17:26.359 up when you're pushing US Congress to adapt stronger data 0:17:26.400 --> 0:17:30.840 privacy laws, don't you think. In California, Governor Gavin Newsom 0:17:31.119 --> 0:17:33.919 signed bills into law that are designed to protect actors 0:17:33.960 --> 0:17:39.200 from predatory practices that involve you guessed it AI. So essentially, 0:17:39.240 --> 0:17:42.480 these laws create rules that media companies are going to 0:17:42.480 --> 0:17:45.000 have to follow if they are to create a digital 0:17:45.040 --> 0:17:48.159 replica of an actor or a performer. One of the 0:17:48.200 --> 0:17:52.359 two laws says that companies must have quote contracts to 0:17:52.440 --> 0:17:55.760 specify the use of AI generated digital replicas of a 0:17:55.800 --> 0:17:59.560 performer's voice or likeness, and the performer must be professionally 0:17:59.600 --> 0:18:04.640 represent negotiating the contract end quote, So no just sneaking 0:18:04.680 --> 0:18:08.920 that in. It has to be more transparent than that. Now. 0:18:08.960 --> 0:18:13.040 The other law requires media companies to first acquire the 0:18:13.080 --> 0:18:18.800 express permission from the estates of deceased actors before media 0:18:18.840 --> 0:18:21.960 companies are allowed to create digital replicas of the dearly 0:18:22.040 --> 0:18:25.080 departed actor in question. So, in other words, you can't 0:18:25.119 --> 0:18:27.840 just go and create a digital replica of Clark Gable 0:18:28.200 --> 0:18:32.680 to be in your movie without first acquiring legal consent 0:18:33.160 --> 0:18:37.879 from Gable's estate. And this relates closely to issues that 0:18:37.960 --> 0:18:41.400 sag Aftra brought up while they were negotiating new contract 0:18:41.400 --> 0:18:44.840 agreements with the movie studios. That's what ultimately led to 0:18:45.480 --> 0:18:49.960 the union strikes. Not that long ago. Have you ever 0:18:50.600 --> 0:18:54.679 been on a social platform like x formally known as 0:18:54.720 --> 0:18:58.840 Twitter and you posted something that you thought was amazingly 0:18:58.880 --> 0:19:03.480 insightful or really funny, or just incredibly relevant to the 0:19:03.520 --> 0:19:06.879 world and what's going on or whatever. Only then you 0:19:07.040 --> 0:19:12.200 absolutely got no engagement whatsoever, you know, no likes or responses, nothing. 0:19:12.680 --> 0:19:15.080 That's a real bummer, right, I mean, here you are 0:19:15.480 --> 0:19:19.639 spitting gold out into the universe and you're getting nothing back. 0:19:20.240 --> 0:19:24.520 Wouldn't it be nice to receive lots of really positive responses. 0:19:24.600 --> 0:19:28.080 Maybe people are riffing on what you said and pointing 0:19:28.080 --> 0:19:32.120 out other things relating to what you were saying, and 0:19:32.560 --> 0:19:37.000 creating a real conversation around it. Well, maybe social ai 0:19:37.160 --> 0:19:41.000 is for you then, though I don't think so. You know, 0:19:41.280 --> 0:19:43.879 social ai is only for you if you do not 0:19:44.080 --> 0:19:47.040 mind that all the responses you get are generated by 0:19:47.040 --> 0:19:51.840 AI bots, because that's what social ai is. It's an 0:19:51.880 --> 0:19:57.960 app that I argue, mimics a social platform because everyone 0:19:58.040 --> 0:20:01.760 else on your instance of this app is a bot, 0:20:02.520 --> 0:20:05.480 Like you're the only human left on Earth and everything 0:20:05.520 --> 0:20:10.200 else is run by AI. You can select the types 0:20:10.240 --> 0:20:13.840 of followers that you'll get responding to your posts. You know, 0:20:13.880 --> 0:20:17.560 maybe you want people who are funny. Maybe you're looking 0:20:17.600 --> 0:20:21.720 for nerds to respond to your stuff. Maybe you're looking 0:20:21.760 --> 0:20:27.080 for insightful observations drawn from what you're posting. I don't 0:20:27.080 --> 0:20:29.359 know how much mileage you'll actually get out of any 0:20:29.400 --> 0:20:31.639 of those, because I've actually been I've been reading some 0:20:31.680 --> 0:20:35.760 of the responses that various reporters have used or have 0:20:35.880 --> 0:20:39.520 posted after they tried out social Ai. They created some 0:20:39.640 --> 0:20:43.000 posts and then they posted the responses they were getting, 0:20:43.359 --> 0:20:47.120 and all those responses strike me as hollow and lacking 0:20:47.200 --> 0:20:50.760 any substance whatsoever. Like to me, it reminds me of 0:20:50.760 --> 0:20:54.160 being in school and you're in English class and you're 0:20:54.200 --> 0:20:56.560 supposed to give a book report, and one by one, 0:20:56.960 --> 0:20:59.399 students are going up to give their book reports, and 0:20:59.440 --> 0:21:02.760 it's very clear who read the book versus someone who 0:21:02.920 --> 0:21:06.000 only read the back book cover, and they're just trying 0:21:06.040 --> 0:21:08.560 to stretch that out long enough to make it seem 0:21:08.600 --> 0:21:11.160 like they read the whole thing. The responses I read 0:21:11.160 --> 0:21:13.320 and social ai made me think of the students who 0:21:13.520 --> 0:21:15.760 just read the back book cover and didn't actually do 0:21:15.880 --> 0:21:18.800 the work. The developer behind social ai is a guy 0:21:18.880 --> 0:21:22.640 named Michael Samon, and I don't know why he made 0:21:22.640 --> 0:21:26.280 social ai. Honestly, I don't know if this was kind 0:21:26.320 --> 0:21:29.320 of his way to lampoon how meaningless a lot of 0:21:29.359 --> 0:21:32.840 online engagement ends up being. Like if you've ever opened 0:21:32.960 --> 0:21:36.240 up threads on Threads or on x or whatever, if 0:21:36.280 --> 0:21:39.919 you just read like the comments section under posts, often 0:21:39.960 --> 0:21:42.560 you come across a lot of stuff that surface level 0:21:42.640 --> 0:21:45.680 is being too kind, there's no depth whatsoever. I don't 0:21:45.680 --> 0:21:48.359 know if Saman was poking fun at Elon Musk for 0:21:48.520 --> 0:21:53.000 buying Twitter and finding out what it's like to have 0:21:53.440 --> 0:21:57.160 an entire platform filled with bots. You know, Musk had 0:21:57.200 --> 0:22:00.639 said before he bought Twitter that Twitter was infested with 0:22:00.720 --> 0:22:02.960 bots and he was going to clean it up. And 0:22:03.160 --> 0:22:05.959 I think a lot of people now argue that X 0:22:06.000 --> 0:22:09.600 has bore bots or a higher concentration of bots than 0:22:09.600 --> 0:22:14.000 Twitter ever had before Musk took over. Whether that's true 0:22:14.119 --> 0:22:17.600 or not, I don't know, but certainly the perception or 0:22:17.680 --> 0:22:19.760 I don't know if Saman was doing this in a 0:22:19.840 --> 0:22:23.880 sincere effort to provide comfort to lonely people who otherwise 0:22:23.920 --> 0:22:27.080 get very little interaction from others. Right Like, if it's 0:22:27.080 --> 0:22:29.520 someone who just feels like they're saying things and no 0:22:29.520 --> 0:22:32.240 one's hearing them, then that can lead to a pretty 0:22:32.320 --> 0:22:36.800 despondent day to day existence. So having something that makes 0:22:36.800 --> 0:22:40.680 you feel heard and seen and validated. That could have 0:22:40.840 --> 0:22:45.040 real value to some people. And maybe that's why Saman 0:22:45.119 --> 0:22:48.040 did this. It's hard to say, because he has given 0:22:48.160 --> 0:22:51.959 various statements that support each of those motivations, and you know, 0:22:52.080 --> 0:22:55.560 maybe he's motivated by a mixture of things, or maybe 0:22:55.600 --> 0:22:58.879 the motivation has actually evolved over time, where maybe maybe 0:22:58.880 --> 0:23:02.640 it started either as a joke or as a sincere 0:23:02.680 --> 0:23:06.120 effort to help people and then slowly evolved into the other. 0:23:06.280 --> 0:23:09.479 I don't know, but whatever the motivation is, I can't 0:23:09.520 --> 0:23:13.240 say that I'm impressed with the quality of the engagement 0:23:13.320 --> 0:23:16.719 you get when you post stuff to social AI. It 0:23:16.800 --> 0:23:20.760 does actually make me think of Threads a lot, because 0:23:20.760 --> 0:23:24.120 whenever I do log into threads, I get the distinct 0:23:24.200 --> 0:23:26.920 feeling that a lot of the accounts that are being 0:23:26.960 --> 0:23:30.720 promoted to me are actually being driven by AI of bots. 0:23:30.800 --> 0:23:33.000 I mean, maybe the account is held by a real person, 0:23:33.040 --> 0:23:35.840 but it's an AI bot that's actually posting the stuff, 0:23:36.040 --> 0:23:39.240 and it's all in an effort to drive engagement. That's 0:23:39.240 --> 0:23:41.720 the feeling I get because there are just too many 0:23:41.760 --> 0:23:45.960 accounts that are all asking essentially the same questions, right, 0:23:46.240 --> 0:23:50.080 and they tend to be questions that prompt a quick response, 0:23:50.200 --> 0:23:54.639 especially if you're linking it to a specific region, like Hi, Austin, Texas, 0:23:54.640 --> 0:23:57.320 what are some great restaurants to look at? Like something 0:23:57.359 --> 0:23:59.840 like that, Right, that's going to get a lot of responses, 0:24:00.000 --> 0:24:02.480 at least in Austin and probably the surrounding areas, and 0:24:02.520 --> 0:24:05.960 it drives a lot of engagement, which, in turn, engagement 0:24:06.800 --> 0:24:10.520 is like currency for influencers. Right. So, I get the 0:24:10.520 --> 0:24:13.080 feeling that a lot of the posts on threads aren't 0:24:13.119 --> 0:24:16.119 being made by actual people. They're made by bots that 0:24:16.160 --> 0:24:19.000 are just trying to get as much engagement as possible, 0:24:19.359 --> 0:24:21.760 and a lot of that stuff ends up being, you know, 0:24:22.200 --> 0:24:25.080 pretty simple tricks, to the point where I very rarely 0:24:25.119 --> 0:24:27.840 respond to anything in Threads unless I actually know the 0:24:27.840 --> 0:24:30.280 person who's posting and I feel like, oh, that really 0:24:30.320 --> 0:24:33.040 came from that person, not this is something that some 0:24:33.200 --> 0:24:37.640 bots spat out in an effort to make number go up. However, 0:24:37.720 --> 0:24:40.760 it's possible I'm just getting paranoid. Maybe I'm just paranoid 0:24:40.760 --> 0:24:43.639 and I believe everybody's a bot and I'm just positioning 0:24:43.640 --> 0:24:46.960 myself to audition for the next version of the thing, 0:24:47.080 --> 0:24:51.240 except it'll be robots, not you. Know Aliens. Maybe that's 0:24:51.280 --> 0:24:54.600 the case. I'm going to take another quick break. When 0:24:54.640 --> 0:24:56.359 I come back, I've got a little bit more tech 0:24:56.400 --> 0:25:09.359