 This 10th year of Daily Tech News show is made possible by you sitting right there walking your dog Thanks to you for including us in your life Dustin Campbell, Tim deputy and Brandon Brooks Coming up on DTNSA City Mayor threatens to sue chat GPT for defamation The US is developing privacy laws one state at a time and why nobody logs out at anything anymore Go ahead try it try to find that log out option. You can't bury Jimmy Hoffa This is the Daily Tech News for Thursday April 6th, 2023 in Los Angeles. I'm Tom Merritt and from studio Redwood I'm Sarah Lane from deep in the heart of Texas. I'm Justin Rabbit. Yeah And on the show's producer Roger Chen. Let us begin the proceedings today As the Lord intended with the quick hits Android developers need to let you delete your personal info or get delisted from the Play Store starting in 2024 Play Store listings must display links in the data deletion section to let users ask for their data or ask That it be deleted this means there must be a web-based option That doesn't require the user to reinstall the app in order to make that request all apps must comply by May 31st 2024 or phase for removal Google also plans to stop apps that offer personal loans from accessing Sensitive user data like images and contacts Carrying on with the Google news CEO Sundar Pichai told the Wall Street Journal that Google plans to let users and I quote Ask questions to Google and engage with LLM's large language models in the context of search We're all shocked Of course, they've said that they would do this before but this is the first time that Anyone from Google has clearly stated it would be a chat interface similar to what Bing does not surprising The first time they said it so I guess it's significant Pichai didn't give any sense of when this would happen though He also said to expect more collaboration within Google's AI units Google brain and deep mind are gonna team up with their brain minds and build something better Microsoft plans to add its Microsoft 365 Copilot feature to one note last month the company announced the AI assistant was coming to teams Word Excel and PowerPoint Copilot will be able to summarize existing notes as well as rewrite or reformat them It can also generate notes based on prompts like make talking points for a meeting for example Copilot for all these apps is being tested in beta and Microsoft has not committed to a final release date The company also added its Bing chat AI to the Swift key keyboard for Android Available by downloading the beta version of Swift key on the Google Play Store. Well, yeah, you don't need to go use the edge browser anymore You just put Swift key on it's kind of kind of brilliant The chips are diversifying folks u.s. Census data shows us chip imports grew 17% on the year in February So we're bringing in more chips building more stuff with chips in them But chip imports from mainland China fell 7.8% where'd they go? Well chips were coming into the u.s. From Vietnam those chip imports rose 74.9% Thailand rose 62.3% South Korea rose 43.3% India rose 34 times what it was a year ago, which is 3400 percent and Cambodia rose 698% those were both coming from very low numbers, but they're in the game They're both in the top 10 now just below Japan The Financial Times reports that advertising on tick-tock grew 11% in March despite concerns over the Security of the app and the potential for the u.s. Government to block tick-tock According to sensor tower ad spending on tick-tock in March was led by Pepsi Door-dash Amazon and Apple among others keep in mind. They may have committed to this ad spending before congressional hearings highlighted Concerns All right, let's talk a little more about This mayor that I mentioned at the top of the surf. Yeah, okay, so His name is Brian Hood. He's the mayor of Hepburn Shire in Australia and sent a letter to open AI Asking it to correct false information being given about him by chat GPT Mayor Hood is considering suing for defamation Members of the public told him that chat GPT had named him as a guilty party in a bribery scandal from the early 2000s Hood notified authorities about the bribes in question and was never charged with a crime He knows about this incident, but chat GPT had it wrong So given that chat GPT is not drawing on facts in this case, but simulating human speech is defamation the right charge here and If so, who is responsible Justin? I'm glad you're here today. So what are your thoughts? Yes, indeed a man who got a b-plus in Kamla at Syracuse University back 1905 I am here to tell you that by the authority of Professor Roy Gutterman Knows, I mean, there's a lot of questions that you have to answer here number one where Case would be tried if he'd be tried in America or in Australia far more Knowledgeed in in American laws, but what you need to do in a case for libel Which is what it's called when it is written when something that is the famatory is written is you have to prove actual Malice that you knew that this was wrong and you wrote it anyway. You published it anyway now on the internet you have a lot of different Things that that protect for free speech or protect for people interacting with their platform Tom can talk to you about section 230 in great length I would wonder whether or not some of that would not come into play here, but I will say that It would be very very very hard to prove based on all of the caveats that open AI lays out for chat GPT that anything was done With malice and that's something that you need to prove in in a legal setting that being said Begin these AI Lawsuits have we are going to see this from copyright to defamation And it's only going to get bigger and bigger and bigger This is a massive sea change in how information is synthesized and that means that we got to write all new rules for it Yeah, I'm looking through my 1989 mass comm law book. I'm seeing nothing nothing about AI in that Everything Justin said was right for the US. It may or may not apply directly to Australia but the the sense of it is still the same which is there is no law about what a System that is trained on facts and also trained on things that aren't facts is Responsible for when it spits out something that isn't a fact and so going back to that actual malice thing Even if Australia has a different way of doing it It's generally the idea behind all defamation law around the world and again I'm no expert either, but the idea is always Did they mean to defame you now some countries occasionally will say like doesn't matter whether you meant it or not Chess EPT is going to be in a different case there, but in all cases. It's not a you It's not a person and so the question of responsibility Becomes a big one which is hey, we said this is not a system that gives you facts We said right up there at the top every time that this gets things wrong Are you responsible for policing it because wow that's going to really change how generative AI can be applied If you're responsible for every single thing it spits out Or is it a case where you're like, yeah, it's almost like a random number generator It's just less random and a lot of its output is more useful Uh just this morning, uh the washington post ran a story Very separate from this But it was somewhat similar a law professor had been named by chat GPT As a sexual harassment offender of a student while on a class trip to alaska That professor says the alaska trip never happened and no one has ever accused me of such a thing And that is a problem. So this is sort of you know, where it gets into The area of well if that happened if that happened to me I want to sue somebody and I want to sue them Yesterday, but who do I sue? I have a question about that though. What's the what's the harm of a law professor Sitting at his desk looked up his buddy found it said something defamatory about him and then told that law professor Hey, this thing said it That's different than chat GPT posted this on google, right, which it didn't do Same thing in right in Hepburnshire one of the constituents said hey, I looked this up and it said this it's wrong Uh, that's different than somebody looked it up on chat GPT and ran a story about it And I think that's where a lot of the discussion when these go to court and when laws are created is going to be around And that's the other part of this is that there's no guarantee that that chat gbt would spit out that same answer twice That's right, and that's going to also play a factor here If you asked it a hundred times and a hundred times it said the same thing Then then maybe that's going to be some standard of proof But but there is and you know the the technology it's why it's important to know about the technology here Right start grafting on other And what what can open ai actually do to say like oh, yeah, we can stop that and what can they be like that's just the way it works and sometimes it's going to say stuff like that and if you know and and Should it not be used at all or should it be like hey check every fact In in its output because it may not be right I wonder if As we go forward and these kinds of cases keep cropping up, you know just saying like hey, this isn't right You know, it's it's like snopes, right? For ai, you know the more humans say oh, no, no, no, let's let's make this better Then you know the the Snopes is snopes for ai because the ai is trained on all the things snopes is correcting as well as trained on snopes Yeah, so it's all in there, right That's yeah, it's uh, it's it's interesting and it's going to keep getting more interesting But it's new right this is this is a new problem that we're just now grappling with of course. There are no laws Thankfully privacy concerns on the internet have been around since the 90s So we're just now getting around to multiple multiple states in the united states passing their own data privacy laws Iowa is the sixth and most recent state to do so joining california, virginia, utah connecticut and colorado These state laws have differences One of the significant ones is that the iowa law requires you to opt out of data collection They're allowed to collect data until you tell them. No, I don't want you to collect it colorado connecticut and virginia have opt in Requirements, but they generally all give you the right to know what data is being collected What it's used for give you a chance to request to have your data exported and even have it deleted You can also allow collection, but not sale like okay if you collect it, but don't sell it to anybody else In some cases and you get the right to move data between services a k data portability in some cases Yeah, Iowa has that provision Utah's law has an interesting spin age verification for certain platforms, particularly social media sites That takes effect march 2024 and oh that's going to be interesting arkansas, ohio connecticut and minnesota are considering similar laws Maybe they should let utah figure out how you're going to actually verify ages When you don't even know if someone is in your state the utah law applies only to people in utah So how does someone determine if someone's really in utah or not without verifying their identity? Lots to consider here varying regimes mean they're going to be lawsuits There's going to be misidentified people ip addresses are not really accurate geolocation Some states are allowing you to sue directly to a company that violates their laws other States require you to complain to the attorney general who then files the case Not to mention the problems with verifying age when you don't know what state the user is from Justin is state-by-state the best way to do this It's the fastest way to begin doing it Because you have more leeway in 50 state houses than you do in our current congressional setting Now i would not be shocked if we saw movement on the federal side on this because as we saw with the tiktok hearings A few weeks ago and the facebook whistleblower last year This there is a moment bubbling about the The effects of social media on young people specifically so i would not be shocked if we saw some kind of federal movement on a An age verification system going forward, but In the meantime it shows the will of the voters if multiple states are pushing things like this going forward Yeah, certainly a will for privacy not all of them doing age verification So some states are fine with kids because social media. Yeah, or at least the current ways that we that we police Yeah, yeah, and then you have senator holly out there pushing a federal ban on social media for for people under a certain age Yeah, and and again, I mean that was something that that we saw From the democratic side with the with the tiktok hearings And and we saw a lot with that facebook whistleblower with with the idea that the social media companies may or may not Almost conspiratorially know the the the harmful effects it has on young people specifically young girls And they do nothing and they continue to weaponize their algorithm to menace the youth Whether or not that is something that we will continue to Have cement and get political traction remains to be seen But the fact that there is movement on this on the state house means that politicians in those states know That there is a positive effect from their electorate if they move forward on it I mean, I guess if you're In the utah case for example, um, utah saying listen, you have to verify your age That is the only way around this and if somebody falsifies that information, then okay, we'll deal with that case by case Oh Gosh, uh, there's so many ways to do that So many ways to do that. I I I would like to see how this Once put in place Is effective and you know how it might Change the way that social networks allow new users to join And and how it how it keeps people off that shouldn't be kept off Because they can't properly meet the requirements for identification. Yeah, that too And considering how much we've talked about bots bots are going to be something that we will talk more and more and more about Sorry to put ai in a non ai story But this is something else that we will see have its effects Maybe this is just part of our new reality as we go forward that for social media hygiene We should be willfully Registering ourselves on these platforms and there will probably be unregulated platforms and then We will we will see which ones take root. Yeah Listen, uh, the you the listener of dts right now Yeah, you kind of missed the days when the news landscape was all about apple, don't you? Well folks know a little more is my podcast the deep dives into individual topics and the first episode of the relaunch is out Right now. It's about rss. It's rocky history how it ended up being the underpinning of podcasting The personalities that fought over it how it doesn't actually stand for anything depending on what time in history you asked If you have not yet sampled the show go do it right now Because big thanks to amos Who edits and produces this and justin at dog and pony show audio? Who came in and up to the game justin? Thank you for for helping us level up with know a little more Absolutely. No little more is a fantastic show and and In my opinion in a world filled with fear uncertainty and doubt as well as a lack of context especially in technology stories I think it's not only a good show It's an important show and it's one that we wanted to make as easy to consume as possible Because we want you to share it with your friends and family I don't think anybody in the business can do it better than tom and uh, I I'm very very Heartened to hear the well wishes on people who have listened to the first episode I can tell you having been working on the next few episodes. You're going to really like where it goes from here Yeah, keep an ear out for algorithm open ai mastodon all coming this season at know a little more dot com All right Quick question. When is the last time you logged out of a website logged out logged out? Yeah, you Maybe you went on uh vacation and you said i'm doing an internet cleanse Or maybe somebody was weird on twitter and you said i screw it shared computer Where you're like, oh, I my bank makes me log out. So that might be the last time that I let forced log out Well, yeah, so banks and you know, health care providers are sort of an exception But for most sites, when are you really logged out? Do you even think about doing that? Do you log into Facebook or twitter or youtube or discord every time you come back to your computer? The answer is probably in many cases. No with all the sites and services that we use the idea of logging out of the internet Kind of doesn't exist anymore. But why? Well, the verge is terry and wane had the same question So spoke to several ui designers in an article titled the impossibility of logging off Uh and in that article it points out that logging out doesn't hold much value for either The company's making the web websites and apps or the consumers that log into them Quote because websites and apps have a profit bearing incentive to keep users logged in and we users We like the convenience of not having to go through the login thing again These days logout buttons are actually buried in settings menus in most cases look at discord not hidden just De-prioritized as something you'd want to do and on mobile devices There are even apps like facebook messenger or whatsapp that don't have logout options But i don't know maybe we want it this way right jesson Yeah, tom there are upsides to this moving from desktop to a mobile device seamlessly as possible because we are always logged on We like the convenience of not logging in constantly or starting from scratch quote-unquote every time we want to interact With an interface or service provider And google meta and countless other companies benefit from this because they can track our habits and learn more about us Since we never really leave Like philosophers in a dark cave. We all saw a different part of this story Uh when we we talked about it sarah came at it from a privacy side Justin came at it from a kind of a futurist bent. Uh, I came at it from an authentication bent So sarah the start with your your privacy analysis would what what did you make of this story? Yeah, so, uh, I Because I hang out at kaiser permanente. Um a medical provider quite often these days. Um, I am logged into that site I don't know Five times a week at minimum But after a certain amount of time like you mentioned justin with banks It'll log you out. That is part of the user interface. That's part of the user experience that is designed to Make the user feel okay. I'm you know, yeah, yeah, I yeah, I'm protected They care about me. Oh good But on something like discord When have I ever logged out of discord? Hmm. What why would I I'm either there or I'm not Maybe I'm on mobile. Maybe I'm on my desktop a variety of different desktops um and I Really don't I don't think too much about that because again the convenience of this Is for the most part during my day, especially because I'm a very online person That makes sense to me, but yes data is being collected about my habits And I I've kind of said like, all right, that's cool But yeah, the whole idea of like going to the internet coming out of the internet We're not we're we don't do that anymore Yeah, and even if you log off sometimes the cookies are still set You know session tracking cookies are still collecting your your your data Justin explain your take on this story So number one I think part of the other element of this is the fact that we've gotten so used to deleting and uninstalling And that has effectively replaced our idea to log out of something we can when we could just eliminate it entirely We probably even feel better than logging out, but this is where I went with it I wonder especially considering we are about to Have much more of a natural language interface with a lot of our technology But by way of AI sorry again for putting a story that didn't have AI, but Just can't get away from it. I know I know I know send all your hate mail to me Or at least my AI system I wonder whether or not There is a world that is coming wherein of course we're always logged into these apps because we are interacting with A natural language model that brings us what we want when we want it And the world of getting the full firehose of twitter the world of getting the full firehose of instagram or facebook Is looked upon as being too much Why do that when you can have your personal assistant bring you all of the relevant Pictures videos tweets and opinions at your demand instead of watching it roll tick by tick by tick by tick by tick And if that's the case it will be predicated on the idea that we are always logged on or We can tell our AI assistant to disconnect from that service. Please close my account or leave it dormant Yeah, I I think I think about it from the authentication standpoint of the reason We don't log out is because it's so painful to log in You have to remember a password and if you're doing it, right? You have a second factor that you have to go dig up and either You know stick into your computer unless you keep your key in your computer all the time or look up a code or get a text message Why would I want to log out if I have to do that all the time? And it is annoying on the things that you should log out of like health sites like bank sites I look forward to the day when authentication gets so secure and fast at the same time Maybe with biometrics some phyto magic that happens in the future that looking at the site logs you in Launching the site. It knows who you are and logs you in and ending the site actually logs you out Closing a browser tab doesn't log you out And but it could and I I could see a I could see a future where we get so good at biometrics and authentication and recognition and computer vision That we can securely log you in immediately because you are the one Using the site and then do the opposite log you off when you're done so that you're not still logged in for whatever reason Anyway, those are our three takes uh on this story. I thought it was a fascinating story because it gave us Such different perspectives among this team. Thank you. Terry Wayne for bringing that up on The Verge Send us your takes feedback at daily tech news show dot com Be real films the makers of my favorite movie not Midsommar we'll begin production this autumn on a tv series telling the story of the pirate bay for sweden's SVT public broadcaster the six-part drama will follow the rise of the bit torrent index its attempts to avoid blocks And eventual decline dynamic television has the worldwide rights distribution rights rather But no release date or platforms have been announced. We don't exactly know where it's going to land yet But dynamic will be looking for deals at mip tv which uh starts on april 17th in can in the south of france I think of all the stories that want to be free The pirate bay story wants to be free For me. Yeah, I mean the folks who did midsommar. So this is this is going to look better than the pirates of silicon valley Uh, I expect well and listen. I said I didn't like the movie. I didn't because it's a disturbing movie But it's well because they were good at it telling the story is why you didn't they were so good at it I watched it twice and said I hate it even more but um, I would yeah, I I feel like uh having covered the pirate bay uh in Various incarnations over the last I don't know. What are we talking about two decades now? Yeah. Um, you know that this is Let's see it Yeah, I'm I'm looking forward to it and I'm jealous that sweden gets it first on on svt They know where they can watch it. The rest of us have to wait. I'm my money's on amazon prime They they're still they're still throwing money around these days. Yeah amazon trying to trying to get that that hot show Yeah My money's on on it just being distributed freely on the internet as it should be Yes All right, let's check out the mail bag Uh tim the dba had thoughts about our ad conversation We were having with scott johnson yesterday and the idea that ai could make ads better Tim says when searching for a product on google I skip past the ads even if they look like a product i'm searching for because it does not benefit me to Be tracked by those ads. I find tv ads very engaging and Choose not to have a tv in my house in large part to avoid being exposed to such ads So do I want to make ads better? Absolutely not I don't want them in the first place and I certainly don't want them to be more effectively influencing me Tim says I agree with sarah that we can't get rid of ads That was the point that I made yesterday tim says do I fall for clickbait from time to time? Of course But I see that as a fault of mine to correct not an activity to improve I'd rather pay directly for quality products like dtns. Thank you tim then be exposed to ads Now I know tim the dba is an exception to what i'm about to say none of this applies to tim the dba, but most humans Do like ads they just don't consider them ads if you don't like an ad It's because it was ineffective because it didn't give you what you wanted now I'm not that's not exactly right There are effective ads that you don't like but then they work because it sticks the brand in your mind and you remember them But most good ads are the ones that you're like. Oh that gave me an important piece of information Thank you for that now. I know about a thing and and so I don't know I I think I think people have attached a moralism to whether ads are good or not That that is that is not exactly fair to the utility of ads even though there's plenty of ads I don't like but that's because they're useless to me and I don't That's all I care about. I want the ads I see to be relevant I would also like to add that while I am positive that tim the dba does want to pay for quality products Including products like this and he does. Yes The question is are you willing to pay the price that the product thinks you should pay or Couldn't you pay if they just mentioned that door dash is a very very good product for which you can get your favorite restaurants delivered to you Use promo code jury Good luck. Good luck with that promo code Well jury speaking of you, um, we uh, we do want people to know more about what you're up to So let let the good people know It's finally here friends The third season of world's greatest con Starring brian brushwood produced by dog and pony show audio It's been a long time coming and it debuts this monday It is an extraordinary season for the first time ever brian and I interviewed First person sources for the story of project alpha if you're not familiar with it You can look it up online when you read the story I guarantee you what you will hear in this podcast is something Materially different for the first time the truth of that will come out and it begins this monday april 10th World's greatest con find it wherever you get your podcast listen to the first two seasons to get ready for the new one Starts this monday Well, uh, yes, put it on your cow everybody Yeah, um, also we would like to put on our new boss for julio Just started backing us on patreon. Thank you for julio. So glad to have you on board Now, I don't know what level virgilio is at Uh, but if it's at the associate producer level, uh, he's now gonna get some some fun little thank yous You know some stickers some postcards I mean if he's up at the master or grand master level he's gonna get some mugs or hoodies every three months As long as as long as virgilio sticks around for three months I'll get one thing every three months all the patrons do all the patrons get access to know a little more Right there in the patreon feed. They get access to the editors desk. They get sarah's live with it They get rogers column, uh, and they get uh the video links to the youtube video of good day internet as well as good day internet itself Uh, and that's gonna happen right now. So patrons like virgilio stick around for the extended show. We're gonna talk about Talking to your ai tired of typing typing means for suckers We're gonna talk about call samantha on good day internet stick around I love my hoodie just saying It was free because i'm a staff member, but they are good hoodies Uh, but just a reminder you can catch the show live money through friday at 4 p.m East urn 200 utc and you can find out more about how to do that at daily tech news show dot com slash live We are back tomorrow talking about alternative fuels for cars with tim stevens and len parolta Drawing the top tech stories talk to you then This show is part of the frog pants network Get more at frog pants dot com Club hopes you have enjoyed this program