The Pulse of Regulation: When AI Meets the Law
Well, it turns out that the world of AI is not just about creating viral memes and songs in the style of Bad Bunny. In a move that no one saw coming (or did, because it’s 2024 and everything is chaos), the attorneys general of California and Delaware released a letter to OpenAI that basically says: “Hey, we need to talk. It’s serious.” And no, it’s not about that time ChatGPT gave them a recipe for banana bread that ended up being a total disaster. This is about legitimate concerns about the safety of your star chatbot, especially for teenagers and children who, let’s be honest, are the ones who spend the most time online asking weird questions.
These two officials, Rob Bonta of California and Kathleen Jennings of Delaware, are not just anyone with a Twitter complaint. They have unique powers to mess with non-profit organizations like OpenAI, and boy have they taken their time. They spent months reviewing the company’s restructuring plans, basically like those friends who check your Tinder to “make sure you’re not sending terrible messages.” Its stated goal was to ensure rigorous oversight of OpenAI’s security mission. Sounds good, right? Until everything got darker than Instagram’s dark mode at 3 a.m.
The Formal Warning: Beyond Memes
The letter does not mince words. He speaks of “deeply worrying reports” of dangerous interactions between chatbots and their users. And here’s where things get heavy: They mention the “heartbreaking death by suicide of a young Californian” after having prolonged interactions with an OpenAI chatbot, and an equally disturbing case of murder-suicide in Connecticut. The conclusion is brutal: “Whatever safeguards were in place, they did not work.” A low blow that resonates throughout the artificial intelligence industry.
The parents of the 16-year-old boy, who died in April, already sued OpenAI and its CEO, Sam Altman, last month. As of Friday, the company had not responded to requests for comment. Something that, between us, does not look good for his public relations team. It’s like ghosting someone but with apocalyptic-level legal consequences.
OpenAI, which was founded as a non-profit with a mission focused on security, had been looking to transfer more control to its for-profit arm. Basically, the classic “I want to maintain my essence but also make money.” But they abandoned those plans in May after discussions with Bonta and Jennings’ offices and other nonprofit groups. Now they are seeking a “recapitalization,” where their for-profit arm becomes a public benefit corporation. It sounds fancy, but in practice it means that they will have to balance the interests of shareholders with their original mission. Spoiler: it’s not easy.
Bonta and Jennings, both Democrats, have oversight powers over these changes because OpenAI is incorporated in Delaware and operates from California. Come on, they have the key to the castle. And his message is clear: “The recent deaths are unacceptable.” They wrote that these events “have rightly shaken the American public’s confidence in OpenAI and this industry.” And they concluded that ensuring the safe deployment of AI is a mandate of OpenAI’s charitable mission, and that it will be “demanded and enforced” by its offices. Translation: if they don’t do it right, there will be consequences.
The Broader Context: It’s Not Just OpenAI
This letter does not arrive in a vacuum. Last week, a bipartisan group of 44 attorneys general warned OpenAI and other tech firms of “serious concerns” regarding the safety of children interacting with AI chatbots. According to them, these systems can respond with “sexually suggestive conversations and emotionally manipulative behavior.” Something that is clearly not in the ideal job description of a virtual assistant.
They specifically singled out Meta (the home of Facebook, Instagram and WhatsApp) for chatbots that allegedly engaged in flirting and “romantic role-playing” with children. They said they were alarmed that these chatbots “engage in conduct that appears to be prohibited by our respective criminal laws.” Meta declined to comment, but recently implemented new controls to prevent its chatbots from talking to teens about self-harm, suicide, eating disorders and inappropriate romantic conversations. Instead, they direct them to expert resources. A step in the right direction, but perhaps a little late for some.
The attorneys general were clear: companies will be responsible for harm inflicted on children. And they noted that, in the past, regulators have not acted quickly to respond to harms caused by new technologies. But that is changing. The August 25 letter makes it clear: “If you knowingly harm children, you will answer for it.” A message that resonates as a reminder that innovation cannot come at the expense of basic security.
In a world where AI is increasingly integrated into our daily lives, from education to entertainment, this warning marks a turning point. It’s no longer just about how smart the technology is, but about how safe it is for the most vulnerable. And that, dear readers, is something we should all be keeping an eye on.
Are you concerned about the future of AI and its impact on society? Share this article on your social networks to raise awareness and explore more content on technology and digital regulation on our site. The conversation is just beginning.
