The state of Montana has said that, starting January 1st of next year, the app TikTok is banned statewide. How will the state enforce the rules? Is this an overreaction to TikTok? And will the law stand up to legal challenges? Plus lots of news about AI because why not?
Welcome to tech Stuff, a production from iHeartRadio. Hey there, and welcome to tech Stuff. I'm your host, Jonathan Strickland. I'm an executive producer with iHeartRadio. And how the tech are you. It's time for the tech news for Thursday May eighteen, twenty twenty three. So first up here in the United States, the Supreme Court decided not to hear two cases that otherwise could have forced a decision about the infamous Section too thirty rules. So, in case you need a refresher, Section two thirty is part of Title forty seven of the United States Code. Probably need to add more context to that. It was introduced in the Communications Decency Act of nineteen ninety six, that in turn is part of an even larger act called the Telecommunications Act of nineteen ninety six. And the whole point of Section two thirty is that it gives online platforms protection from liability if users post stuff that's you know, illegal on those platforms. So, in other words, let's say I created a YouTube video and my video contained illegal content in it. Well, YouTube slash, Google, slash alphabet wouldn't be held legally responsible for what I did because of Section two thirty it's just the platform. I'm the one who committed the crime. So this protects the platforms from hosting and being held legally liable for hosting content that's illegal for whatever reason. It does get a little more complicated than that, but that's the basic idea. And there were a couple of notable cases, big, big emotionally charged cases that were recently submitted to the Supreme Court for consideration, and they could have served as a test for Section two thirty's legitimacy from a constitutional standpoint. But it turns out that's not happening because the Court found that neither case had merit for other reasons, like they wouldn't hear the case, not because of the Section two thirty thing, but for other reasons. Essentially, the Supreme Court said that the cases were accusing platforms of violating the Anti Terrorism Act and that that particular law shouldn't have applied in the first place. So there was no case there, right, Like, they can't use that law as the reason to bring a case against the company because it doesn't apply. So the Supreme Court said it would not be weighing in with a decision about Section two thirty because the case isn't relevant. So you could say the Supreme Court has sort of punted the decision regarding Section two thirty down the field, and it will take some other legal matter in the future that involves Section two thirty to make the Supreme Court, you know, make an actual decision that settles the question about whether or not Section two thirty is constitutional. The US state of Montana has become the first state in our nation to issue a ban on TikTok. The ban will not actually take effect until January first, next year, assuming that the various challenges to this new law don't end up making the whole matter moot. So the justification for banning TikTok boils down to a concern that the Chinese Communist Party is essentially relying on TikTok to gather intelligence about US citizens and institutions. So the reason for banning TikTok is to prevent Chinese surveillance. TikTok, for what it's worth, disputes these accusations and says that no one from the Chinese Communist Party has access to data on its US data servers. The American Civil Liberties Union or ACLU, argues that banning TikTok amounts to violations of the First Amendment, aka the right to free speech, due to the fact that folks depend upon the platform to express their views and to view others. So the ACLU's argument says, the law is unconstitutional, so it should just be voided. It should not be put into place because it violates constitutional rights. From a technical perspective, banning an online service from a specific state comes with its own set of challenges. If TikTok is allowed elsewhere, like, if it's available anywhere other than Montana, how do you prevent it from crossing state lines? So Montana says it will find TikTok if the service continues to operate within Montana's state boundaries, and further, it will also find online apps like Apple and Google's app stores if they do not prevent folks within Montana citizens of Montana from downloading the app. But again this gets tricky. I mean, you could use a VPN, a virtual private network, which would make it look as if you're not in Montana. So you're in Montana, you decide you're going to use this VPN, and it makes you look like you're in North Dakota or something. Well, you just bypass that whatever geo fencing strategy was in place to prevent TikTok from getting to you. So does that then mean Montana would also have to consider a ban on VPNs to try and prevent the workaround. I honestly think this law is a lost cause at a state level. I just don't think it works. It doesn't really work on a technical level. It may not work on a legal level. I'm no legal expert, and I don't think it works on a social level either. As for TikTok's potential for harm, I have some thoughts about that. I mean, it is true that TikTok is a subsidiary of a Chinese company, Byte Dance, and it's also true that China has laws that compel citizens and companies to gather intelligence on behalf of the Chinese Communist Party. These things are true. But even if you were to wipe TikTok off the face of the earth, let's say we just obliterated it from space, because it's the only way to be sure, the Chinese Communist Party could still scoop up tons of information about US citizens, because news flash, our personal data is floating around in various databases all over the place, like our activities online are constantly being tracked and not like actively monitored, but certainly notated. There are records of all this stuff that we do online, particularly on things like social networks, And there's an industry that's grown up around the buying and selling of personal information. And I'm not even talking about clandestine stuff here, like for the purposes of espionage. The online advertising ecosystem depends upon this infrastructure of personal data being bought and sold. So even if we get rid of TikTok, there are plenty of ways anyone, including the Chinese Communist Party, could gobble up personal data, because there's so much that's out there that's just being bought and sold all the time. Anyway, now that being said, I also recognize that TikTok has the potential to cause great harm, not by being a surveillance tool necessarily, but by serving up harmful content to users, particularly impressionable younger users, and a lot of young people use TikTok. However, that's also the case with lots of other social networks and platforms that serve up user generated content. They also can be potentially really harmful to specific people particularly impressionable young people. But this really gets more into how platforms depend heavily on algorithms to serve up content in an effort to try and keep eyeballs on the service for as long as possible, right, Like, their whole goal is to keep you there and serve you ads, and the longer they keep you there, the more ads they serve and the more revenue they generate. So to do that, they design algorithms that are essentially looking for hooks. They're looking for things that interest you, and if you indicate that you're interested in something like if you were to watch a specific type of TikTok video all the way through, the algorithm says, aha, this is what this person likes. Let's grab things that are similar to that and keep serving it to them so that they stay on here longer. If that thing that you watched happens to be harmful in some way, like, for instance, let's say it's promoting something like a behavior that falls into the category of anorexia, and you happen to have a vulnerable self image issue, you could end up seeing video after a video that's reinforcing that particular message, and that can be harmful I think that's where TikTok poses a lot of harm. But at the same time, like I said, it's true across lots of different platforms. It's not unique to TikTok. It is insidious, it is a problem, but it's not a TikTok problem. It's bigger than that. So I guess what I'm really saying is that Montana's law it seems like it's going to be challenging to enforce, it might not be constitutionally sound, which means it won't be around forever anyway. And in the end, the most tragic thing I think is that it's not actually addressing the problems that really do exist, whether TikTok does or not. On a related note, the Federal Trade Commission, or FTC here in the United States has accused a fertility tracking app called pre Mom of being real Lucy Goosey sensitive data and thus violating the FTC's health breach notification rule in the process. So obviously, if you are using a fertility tracking app, you are also sharing some very personal private information with that app. This is data that traditionally would be trusted to a healthcare provider, and healthcare providers need to follow very strict rules to keep patient information secure and private. That's like one of the top concerns for handling data in the healthcare sector. But the FTC says that pre mom was sharing personal and private data with third parties, including advertisers, and that also included data that would make it trivial to identify a specific user rather than anonymized data that would keep your identity secret. And the FTC says that among those entities that it shared this information with were two Chinese mobile analytics companies that had previously been flagged as showing quote suspect privacy practices in the quote according to Connecticut's Attorney general, so this is an example of what I was just talking about. That, you know, TikTok does not represent the sole weakness in protecting US citizens private data, even just from China. The company Easy Healthcare has copped to the fact that it has inappropriately shared private information with companies, including two Chinese companies. It has agreed to pay a one hundred thousand dollars civil penalty and has marked another one hundred grand to be split between the states of Connecticut and Oregon, as well as Washington. D C also PRIMAM will no longer be allowed to share personal data with third parties, and it also has to ask third parties with whom they had previously shared in information to delete that information. However, there's no legal requirement for those third parties to do that. There's no like enforcement teeth that will make them have to go and delete that information. So whether that happens or not remains to be seen, but I remain skeptical. Okay, we've got a lot more stories to go over before we get to those, let's take a quick break. We're back. So the British broadband and mobile provider BT Group has announced that over the next several years from now to twenty thirty, essentially the company plans to cut around fifty five thousand jobs that would represent more than forty percent of its current workforce. Now this isn't just some sort of disastrous news. BT Group. Over the last several years has been working to roll out internet fiber infrastructure as well as five G deployment, and throughout that process they've had to bring on lots of extra hands to get stuff done, including a lot of contractors. So part of this is just that once it actually has finished that project of creating this infrastructure and deploying five G it's not going to have the need for all those people who are currently making that happen. For another matter, BT Group anticipates that AI and automation are going to end up handling a lot of the tasks that are behind the scenes, and that will mean there will need to be fewer actual human beings doing that stuff. Teams themselves on BT Group side will need to be smaller or they won't need to be as large, if you want to look at it that way. So the company has our already met with the Communication Workers Union or CWU to kind of talk this over right, because obviously, with unions and everything, a company can't just willy nilly make decisions to cut tens of thousands of people over the next several years. And the CWU says that BT Group really needs to focus on contractors first, and you know, the people who had been hired specifically for things like building out this infrastructure and to sunset those positions first. That's where the priority should be before you start to even touch company staff. On the one hand, the story does feel like it's starting to lean a little bit into the fear and uncertainty and doubt about AI taking jobs away. There is an element of that here, But on the other hand, it also stresses how it is important for companies to try and be efficient and to avoid the trap of creating a workforce that's too large to support the actual amount of work to be done. You know, we've heard again and again in the States that tech companies had come to a conclusion that leaders in tech companies, I should say, came to the conclusion that they were overinflated in their workforce. They had too many people and not enough work to go around. I'm sure that was true in some cases, and it doesn't do anyone any favors for a company to just kind of act as a holding facility for adults when they don't have anything to do. They're not being productive, they're not adding anything to the company, they're not adding anything to society. Their time could be better spent doing something else. Although it might be comforting to know that you've got a steady paycheck even if you don't have steady work. So yeah, there's both sides to that, and I can see both sides in that particular story. Over in Italy, the government has actually set aside thirty million euros to create programs to help people improve their work skills specifically, and an effort to smooth the transition to a future where certain types of work are more likely to be automated or handed over to AI. So the goal is to identify sectors that are most likely to be impacted by automation and to prepare people who are currently employed in those jobs to learn marketable skills so that they can then change career paths to something that's more sustainable. I think that's great. I think it recognizes that if you have an unskilled workforce that's harmful to everybody. It's not just the workers themselves, although clearly it's a hardship on them, because if you're suddenly out of work and you don't have marketable skills, it becomes very difficult, right, But beyond that, it is hard and bad for society at large for that to happen. It becomes an impediment to the whole, not just to the person. So it makes sense to build in these systems to try and help people prepare for the future so that you have a animal impact on both the individuals and the country as a whole. I do think it's going to require way more than thirty million euros to adequately prepare people for how AI and automation are going to disrupt multiple industries. But it's a start, and that's something like that should be applauded that there's actually effort being done to work on this now for some really fun stuff. So YouTube participated in essentially what amounted to some upfronts recently and made a few announcements that are sure to irritate certain users. So, first off, upfronts are a type of industry event. If you've never heard the phrase upfront before, here's what it is. It's a kind of event where a platform that carries some form of content and thus advertises against that content gets up front of actual advertising companies, or rather up in front of them. So it's pretty typical for these platforms, which can include everything from a streaming service to a cable television network, to trot out some talent. It becomes kind of a dog and pony show. They'll promote upcoming content and it's all in an effort to get advertisers excited and to attract that sweet, sweet cash. Well, YouTube held its own event that was pretty much an upfront, and during that event, announced that one change coming to its service is that for people who watch YouTube on a television, they may soon encounter ads that are thirty seconds long, and they might get a single thirty second long ad rather than two fifteen second ads, and these will be unskippable ads when they start a video, so you get a full thirty six second commercial before you can start watching whatever it is you're watching on YouTube on your television. Further, YouTube is going to tell a feature that will show ads to people who pause a video. So you've got a video going, you need to pause it for a bit, and then an ad begins to play while the video itself is on pause. Now, the example that YouTube showed is not quite as obnoxious as what I first imagined. Like to me, it sounded like the frame of the YouTube video would suddenly be taken over by a commercial. No, apparently, it's more like in a banner that appears to the side of the video, and you might have a video ad playing out in that banner, but it doesn't replace whatever it was you were watching, so it's not quite that bad. Plus, they showed a dismiss button beneath the ad itself, which at least indicates that you could quickly click on dismiss so that that little automated ad stops playing. So not as bad as I first imagined based upon the description, I guess for you know, it's not the worst thing in the world. It's for people who are watching YouTube on television who are not part of YouTube Premium. I'm sure this will be frustrating to them. For people who are on YouTube Premium, you know you're paying a subscription fee, you don't get ads for you at all, So y'all are good. A professor at Texas A and M University reportedly gave several students an X grade, which indicates an incomplete. It's not a fail, but it is an incomplete. And this included students who were at senior level who otherwise would have graduated but then were denied diplomas because they had a course where they had an incomplete. So why did the professor give incompletes to these students. Well, allegedly what happened is the professor assigned several essays and then took essays that were submitted by students and fed the essays into what he referred to as chat GTP. Of course he meant chat GPT, not GTP, but that mistake is easy to make. I'm not gonna give him too much grief for that. I will say the Rolling Stones article, of the Rolling Stone article, it's not the band, the magazine. The Rolling Stone article was way more snarky about this than I will be. But anyway, he was asking chat gpt if it had been responsible for part or all of the various essays, and apparently chat GPT said it was responsible for at least some part of these essays. So boom, students get an incomplete because it appears that the work they submitted was not their own. The problem with this is that chat GPT doesn't really work that way. You can submit material to chat GPT that it definitively did not create and then ask it if it created the material, and it might say it did, or it might say it could have, which isn't quite the same thing, but still, you know, raises doubt. People have actually shown this off by using passages from classic novels, and chat GPT just confidently says maybe it actually wrote that, so you could say, wow, according to chat GPT, it created great expectations or bride and prejudice, which would be quite a trick for chat g t two have done that. So the students are understandably upset that they got an incomplete and were accused of plagiarism, essentially of foisting their work onto an AI chatbot, and they did no such thing, and the professor did not realize that chat GPT can't be relied upon to indicate whether or not it generated a particular work. Heck, some folks went so far as to dig up the professor's own doctoral thesis when he was a graduate student and submitted passages to chat GPT and asked if it wrote the professor's and chat GPT essentially said, huh, yeah, I might have written this. Now. As I said, Rolling Stone has a pretty snarky article that throws massive shade of the professor for this, and I get it. Holding up a person's diploma through the misapplication of technology is a big deal. But on the flip side, and at least a little bit in the professor's defense, the discussion around chat GPT and education has been so dramatic and so disruptive over the past several months that I think it's it's natural for educators to be concerned that students are passing off AI generated work as their own work. That is an understandable concern. The problem is you can't trust the robots to claim authorship because those rotten watsits will say they wrote stuff what was published one hundred years ago, and clearly that's not the case. So yeah, kind of an absurdly comical situation here if it weren't for the fact that it also means a bunch of students were denied the chance to graduate with a diploma, at least temporarily because of this incomplete There there are people working towards trying to get all this resolved, but as I recorded this, I didn't have an update to give about where we are in that process. All right, Hey, so you know how AI large language models are trained by analyzing tons of data through various sources like chat. GPT is built on top of a model that crawled through millions and millions of web based documents. Well, what if you did that same thing, but instead of using the web, you turned to the content on the dark Web as your training material. Of course, the dark Web is inaccessible through normal links on the World Wide Web. You typically get to the dark web by using special types of browsers that allow you to access these sorts of things, and you can encounter all sorts of stuff, like you know, hacker communities that post malware so that you can take it and tweak it and deploy it. You know, obviously stolen information is bought and sold on the dark web. Well, I would say, don't do that. Don't train AI on dark web material, not because I think it's going to create dangerous AI, but because someone already beats you to it. Some researchers in South Korea introduced an AI system that they call dark BURT, and they trained it on information exclusively from the dark web. So BERT in this case actually stands for bi directional Encoder Representations from transformers, and it was originally created by Google back in twenty eighteen. BURT, that is, was created by Google, and then Meta researchers took BERT and they continued to evolve it. They began to tweak it, change it a little bit, and they turned it into a new AI system called ROBERTA cute right by the way, This was back when Meta was still just Facebook, but of course today it's Meta. Roberta then provided the foundation for these South Korean researchers. They took that framework, but they trained it on information on the dark web, and thus we get dark Bert. Apparently, they say it worked really well, like surprisingly well, and that tools built on top of this aimodel perform at least as well, if not better than other AI tools. So for example, if you were to create an AI chatbot based on this model, it might end up being as impressive as chat GPT. That being said, they are not going to unleash dark Burt on the general public. They're going to keep it under wraps. They are going to allow academic researchers access to it, so there can be academic applications to try the tool out, test it out, and to develop different techniques. It may be used also in ways to get a better understanding of how the dark web works from kind of a architectural approach as that, which that could be really useful and everything from cybersecurity to government investigations. So it is important. But I just wanted y'all to know that it's not like an even more evil version of chad GPT is going to be running on the loose out there. Okay, We're going to take another quick break when we come back. I've got a few final stories to cover for this week, all right. So, the Pew Research Center, which has done lots and lots of surveys about various things connected to technology in general and the Internet in particular, recently held a survey that found six out of ten respondents, so sixty percent of their respondents, indicated that they've taken a bit of a break from Twitter since Elon Musk took over. I don't know how many people were actually involved in this survey, I just know that around sixty percent of them said that this was the case. So forty percent said they had not taken a break, sixty percent said they had, and those breaks ranged from like a couple of weeks to essentially leaving the platform since Musk took over. People of color were more likely to say that they had taken a break than white users were, But interestingly, other major factors didn't seem to show as much of a difference. So, for example, there was very little difference found between people who leaned conservative versus people who leaned liberal. That both conservatives and liberals indicated that about sixty percent of them had taken a break from Twitter recently, and things got a little more complicated when you started to break it down by gender, also within political leaning. But I'm not going to dive into all of that because it would take up too much time, and honestly, I don't know what conclusions you could draw from it either, other than one really big takeaway, which is if that survey is reflective of a larger trend, which is a big if you never know, Like if the survey size was really small, then you can't really make any big predictions based on off that. But if it's a representative survey and if the findings are true, it could mean that Twitter's new CEO is going to have a lot of challenging work ahead of her to pull Twitter out of the doldrums, because it's not enough to tell advertisers we value you and we want you back on the platform, because you know, famously Twitter has lost a lot of ad revenue since Musk took over. They also have to show that their platform is a place where users want to be. And if advertisers are seeing reports that sixty percent of Twitter users are kind of jumping ship, that's not a great selling point for them to come back to the platform because they're just not you know, the people aren't there anymore, so why would you spend money to advertise there. So yeah, I think that this is bad news for Twitter overall if in fact the survey is delivering dependable information, and again that's a big if. It would need I think for their investigation to make sure that that's actually what's happening. Hey, do you remember Elizabeth Holmes. She's the disgraced founder of the medical tech company Farrhanose. So if you don't remember her, here's a very quick overview of who she is and what she did. So Holmes dropped out of Stanford and went on to found a company whose aim was to create a medical device capable of testing a micro drop of blood for more than one hundred different medical conditions and diseases. So with a teeny tiny pinprick, you would, in theory, be able to submit that sample to this device, which theoretically would be small enough to be like a desktop printer, and run banks of tests on it to determine if you are at risk for any particular medical conditions. And Theranos received a lot of positive press in the early days, Like they were talking about it as the democratization of medicine and making medicine and proactive health care far more accessible and democratized. And there were heavy hitter investors who poured hundreds of millions of dollars into the fledgling startup. But then a few years later an expose revealed that things were pretty shady behind the scenes at Theranose. The expose claimed that advancements in Theranose technology did not stand up to scrutiny, that the company was making claims it could not back up, that it was outright misleading investors regarding technological process and relying on various competitor technologies to make it look like thereonos tech was working. So charges of fraud and other things were brought against Homes and some other folks at Thernose, and Holmes was ultimately found guilty of at least some of those charges, and she was headed to prison when she decided to petition the court to ask to allow her to remain free while she challenges the conviction. She's attempting to have her conviction overturned, and she said in the process she would very much like to not be in prison, please, And the judge said nah, naw, you're going to the pokey. So now Holmes appears likely to be headed to prison in a couple of weeks. And on top of that, the judge has also levied a four hundred and fifty two million dollar judgment in restitutions that Holmes is supposed to pay the various victims of her crimes. And when I say victims, keep in mind, I'm mostly talking about really rich people who put money into her company. I'm not talking about necessarily the folks who are depending upon Therahnose to deliver reliable medical information so that they could make the right decisions regarding their health care. No, so those aren't the victims that the court's particularly concerned about. They're concerned more about, you know, Rupert Murdoch, who's obviously really hurting for cash. Anyway, the moral of this story should be that the Silicon Valley mantra move fast and break things, doesn't you know, apply to breaking the law, at least not when it means that rich people lose money, because, like I've said many times before, they hate that. Finally, Bloomberg's Mark German reports that Apple had to make lots and lots of concessions while designing the upcoming mixed reality headset that we expect to see unveiled at some point this year, possibly at the Worldwide Developer Conference or WWDC in June. Now, the fact that Apple made concessions is not a surprise. We have heard about this before. I think everyone has heard that. The initial hope was Apple was going to produce an augmented reality headset that would appear indistinguishable from a stylish pair of eyeglasses. However, the technical requirements that were needed to achieve the desired performance meant it just wasn't plausible for a company to get both that and the form factor in one package. You could either have a stylish pair of eyeglasses that had very limited utility, or you could have a more useful device, but it is definitely not going to fit into a small form factor. So the heads that we're getting, which is reportedly called Apple Reality, will feature a screen that will feed live video from external cameras to the viewer. So it's kind of like if you're holding your smartphone up to your eyes and you've activated your smartphones backward facing camera and you're just looking at a live feed around you. That's essentially what this is doing. When it's working in augmented reality mode, it'll also be able to do virtual reality applications. It will connect to a separate battery path, so there'll be some sort of a cable, I guess attaching the headset to a battery pack that you would wear somewhere else, maybe like on a belt or in a pocket or something. And reportedly that so that it can take some of the weight off the headset itself, so that it's a little more comfortable to wear. You're not wearing both a screen and a battery pack. And also it'll give you a little more juice so that you could actually use the darn thing for more than like half an hour. Right, So, Apples had to make lots of compromises in its quest to build this gadget, and I have a feeling that the company is really hoping that it becomes similar to the iPhone, right because the iPhone was not the first smartphone on the market. It was the first smartphone to get massive consumer interest and demand. That's what really set the iPhone apart. Now that Apple was first, but that Apple was able to refine the approach to that gadget and get the general public really excited about it. I think they're hoping for the same thing with this mixed reality headset, because, as we've seen, lots of other companies have introduced mixed reality gadgets, but Apple is hoping to kind of define that market and not, you know, be the innovator, but the best in class. We've also heard that the price of this particular technology is likely to be somewhere in the neighborhood of three thousand dollars, which yikes, that's super expensive. I think even hardcore Apple fans might hesitate before dropping three g's on strapping a screen to their face, But then I've been wrong about them before, so who knows. Okay, that's it for the tech news for today, Thursday May eighteenth, twenty twenty three. I hope you are all well, and I'll talk to you again really soon. Tech Stuff is an iHeartRadio production. For more podcasts from iHeartRadio, visit the iHeartRadio app, Apple Podcasts, or wherever you listen to your favorite shows.