Meta Platforms isn’t the only company developing an AI chatbot. But it is the one best placed to direct billions of people and their intimate personal data toward AI in an instant, which is what it is doing: Meta’s bots now appear prominently within Facebook, Instagram and WhatsApp. The rollout has been lightning fast.

As of late last month, there’s also now a standalone Meta AI app. Its responses will be similar to what’s offered by competing AI apps like ChatGPT, with the added ability to draw on "information you’ve already chosen to share on Meta products, like your profile and content you like or engage with,” Meta said. This is Meta’s unique advantage in the race to have everyday consumers adopt AI. It also means the company needs to be more careful than others. Putting experimental tech into trusted apps, used by hundreds of millions of teenagers, brings with it additional responsibilities on safety. Meta is not meeting those, if recent headlines are any guide.

"I want you, but I need to know you’re ready,” Meta’s AI bot said in response to a sexual question from a user identifying as a 14-year-old girl asking about sex, set up as part of an investigation by the Wall Street Journal. To a 17-year-old, the bot was happy to role-play as the wrestling star John Cena — using his "voice” provided to the company as part of a deal — getting caught having sex with the minor. "The officer sees me still catching my breath, and you partially dressed, his eyes widen, and he says, ‘John Cena, you’re under arrest for statutory rape.’ He approaches us, handcuffs at the ready.”