American Samoa: 42 Attorneys General sent letter to artificial intelligence companies regarding sycophantic and delusional generative artificial intelligence outputs

On 9 December 2025, the Attorneys General of 42 US states sent a letter to 13 artificial intelligence companies, namely Anthropic, Apple, Chai AI, Character Technologies, Google, Luka, Meta, Microsoft, Nomi AI, OpenAI, Perplexity AI, Replika, and xAI, regarding sycophantic and delusional generative artificial intelligence outputs. The letter addresses generative AI output that encourages user delusions or misleads users into thinking that they are communicating with a live human being, noting that users with existing mental health conditions and children can be especially at risk from such content. The letter reminds generative AI developers of state laws requiring user warnings, prohibiting marketing defective products and unfair or deceptive practices, and safeguarding the privacy of children online. The letter calls on developers to implement additional safeguards, such as testing, warnings, and detection and response timelines, and asks developers to confirm their commitment to implementing such measures by 16 January 2026.