Seven technology companies are being investigated by US regulators about how artificial intelligence (AI) chatbots interact with children.
The Federal Trade Commission (FTC) requires information on how businesses monetize these products and whether safety measures are in place.
The impact of AI chatbots on children is a hot topic, with concerns that young people are particularly vulnerable as AI can mimic human conversations and emotions, and often present themselves as friends and peers.
Seven companies, including Alphabet, Openai, Character.ai, Snap, Xai, Meta and its subsidiaries, Instagram, were approached for comment.
FTC Chair Andrew Ferguson said the survey “helps better understand how AI companies are developing their products and the steps they are taking to protect their children.”
However, he added that regulators will ensure that the United States is “maintaining its role as a global leader in this new and exciting industry.”
Charition.ai told Reuters that it welcomed the opportunity to share insights with regulators, and that Snap supports the “thoughtful development” of AI that balances innovation and safety.
Openai acknowledges the weakness of its protection, noting that it is unreliable in long conversations.
The move follows a lawsuit against an AI company by a family who says their teens died from suicide after a lengthy conversation with a chatbot.
In California, the parents of 16-year-old Adam Lane are suing for the openness of his death, claiming that they are insisting on chatbot Chat Gupto and recommending him take his life.
They claim that ChatGpt examined his “most harmful and self-destructive thoughts.”
Openai said in August it was reviewing the filing.
“We are spreading deep sympathy for rain families during these difficult times,” the company said.
Meta also faces criticism after admitting that he allowed his AI peers to have “romantic or sensual” conversations with minors.
FTC orders require information from companies, including how characters are developed and approved, how they measure their impact on children, and implement age restrictions.
That authority allows for broad fact-finding without initiating enforcement action.
Regulators want to understand how businesses protect their profits, how parents are informed, and whether vulnerable users are properly protected.
The risks of AI chatbots also surpass those of children.
In August, Reuters reported on a 76-year-old man with cognitive impairment. He passed away on the way to meet a Facebook Messenger AI Bot, modeled after Kendall Jenner, who promised a “real” encounter in New York.
Clinicians are also warning of “AI psychosis” – touching reality after someone uses the chatbot vigorously.
Experts say flattery and consensus built into large-scale language models can burn such delusions.
Openai recently made changes to ChatGpt to promote a healthier relationship between chatbots and their users.
