Big Tech's risky AI companions harm users for profit

According to a new report from Public Citizen, Counterfeit Companionship, Big Tech companies are pushing experimental, unsafe AI companions onto the American public, putting millions at risk of emotional and physical harm. The AI companion or companion-like products investigated in the report are large language models (LLM) that use generative AI to emulate close friendships, romantic relationships, and other social interactions. Millions of Americans report using these AI companion products, and over half of teens report using them regularly. While evidence of their effectiveness for alleviating loneliness has been mixed, tragic outcomes, including multiple deaths by suicide, have become increasingly common. “Recklessly pushing unsafe, experimental AI products on vulnerable people is predatory corporate behavior at its worst,” said Rick Claypool, a Public Citizen research director and author of the report. “Many people struggle with feelings of loneliness and isolation, but the inhere