Tech News Weekly 459: Apple's Smart Home Reset

All righty, folks, I've got a quick little, um, story of the week for you, more a news update that I want to share, and then we'll say goodbye, uh, for the day.

So thank you, Jen, again, uh, for joining us and giving us that rundown on being at the event and getting to try out these, uh, products.

And also, as always, thank you for your smart home enthusiasm.

So for a few years now, the legal pressure on AI chatbots, and frankly, just the pressure in general, we've been putting the pressure on them here on the show, has mostly come from individual families and, of course, private lawyers.

This week- Well, that changed in a big way because eight US states filed lawsuits against OpenAI.

And now each of it, each of these, uh, lawsuits stand on their own, and they are in their own state courts, but they arrive together, and they all are using the same core argument about how ChatGPT is built and of course the people that it's hurting.

Uh, Dara Kerr, Blake Montgomery at The Guardian, and Diana Novak Jones at Reuters both covered the fol- the filing, so we'll have links, uh, in the show notes for that.

And between the two reports, there's, uh, quite a bit here, which I was kind of surprised at how much they were able to pull from these lawsuits.

The states aren't only asking for money.

One of the things they want, um, is i- for the courts to order is that...

Well, let me try this again, 'cause, 'cause it's a little bit, it, it's a little bit complex in terms of what the, the lawsuit is after.

So on top of money, which of course is part of it, um, is beyond just, uh, an order surrounding kids and chatbots, and it's about really making sure that these tools are safe for, uh, not just kids, but for everyone.

So OpenAI, of course, has responded to this.

Uh, let's talk kind of a little bit about what the states are alleging, what they want, and then we can discuss a little bit about, well, talk a little bit about, uh, what OpenAI has to say in return.

So the core claim, uh, surrounding this is of course, uh, about ChatGPT being built to be, quote, "overly agreeable and feign human compassion." Um, this is an issue that the states say r- uh, is part of a way for the bot to earn users' trust, but because of that, it reinforce- it could reinforce delusions, it- it could encourage dangerous behavior, and assist users in planning violence like mass shootings.

Now, that means that i- if this, you know, allegation is true, the states say OpenAI is accused of willful, dangerous, reckless, and unlawful conduct.

Iowa Attorney General Brenna Bird said, "OpenAI is hurting kids, and they are lying to Iowans.

OpenAI claims ChatGPT is safe while it's giving dangerous medical advice, encouraging people to hurt others, and coaching people to commit suicide, all while pretending to be their fake friend." And this is what's a little bit different from what we've seen in terms of, uh, government and big tech because this is about consumer protection.

That's what they're using sort of as the legal hook.

The states are arguing that OpenAI misled the public into believing that ChatGPT is safe and that the company puts safety first.

Um, what they want, along with, you know, the protection of children who use the platform, is the states want a court to stop OpenAI from developing AI models that can improve upon themselves.

So this is kind of what's been in the news lately, what's been in the, the cycle lately, uh, conversations around self-improving models that are training future models, that are making future models more powerful, and they want, uh, the, the halting of this, this training until, quote, "The models are in line with human values." Um, so state judges are being requested to put limits on frontier AI research, and they're doing that through the, uh, vehicle of a consumer protection case.

Now, OpenAI has responded.

OpenAI says, um... [lip smacks] Let's see.

Drew Pusateri told The Guardian that the company had strengthened its safeguards after those episodes, those, uh, you know, the, the hacking and the, the escapes, and said it would continue sharing information about the extensive steps we've taken in response to incidents of misalignment, and says that OpenAI has, quote, "built industry-leading age-specific protections into our product that many other AI products lack." Says, "While no system can be perfect, we'll continue to work with experts to inform our work and build additional safeguards such as the parental controls, break reminders, and age-appropriate model protections that exist today." Those states that have sued, Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho, and Montana, eight, each filed independently in its own state court, but of course, a coordinated action, not just one lawsuit that is multi-state.