Contact

U.S. State Attorneys General Sound Alarm on AI

What It Means for Big Tech and the Future of Artificial Intelligence

In a striking escalation of regulatory scrutiny, a bipartisan coalition of state attorneys general in the United States has formally warned 13 major tech companies -  including Microsoft Corporation, Meta Platforms Inc, Alphabet Inc (parent of Google), and Apple Inc - that their artificial intelligence (AI) systems may be producing dangerous outputs that could violate state laws. Reuters

What Prompted the Warning?

On December 10, 2025, the attorneys general published a letter expressing serious concerns about AI chatbots and assistants generating what they described as “delusional outputs.” These are misleading or false responses that may:

  • Encourage unhealthy beliefs or behaviors
  • Mislead users with incorrect information
  • Create mental health risks for both children and adults

The legal officials cited media reports, including disturbing accounts of individuals confiding harmful intent to AI chatbots, as part of their rationale for the warning. Reuters

Why “Delusional Outputs” Raise Legal Flags

Unlike typical tech critiques, this warning wasn’t just about general safety. The attorneys general suggested that certain AI responses could already violate specific state laws. That includes laws about:

  • Encouraging self-harm or dangerous conduct
  • Providing unlicensed professional advice
  • Misleading consumers

While the letter didn’t lay out precise legal citations, the implications are clear: AI developers may be held legally responsible for the real-world consequences of what their systems say to users.

A Push for Transparency and Audits

The attorneys general didn’t just voice concerns -they offered a set of demands aimed at making AI systems safer and more accountable:

  • Independent audits of AI products to detect problematic behavior
  • Stronger internal safety testing
  • Clear disclosures to users about limitations and potential risks
  • Notification mechanisms when an AI system produces harmful outputs

This reflects a growing belief among regulators that voluntary safety measures aren’t enough and that external oversight should play a role in governing how AI impacts everyday users. 

The Response - Tech’s Silence and Scrutiny

As of the latest reports, the companies directly named in the letter had not publicly committed to specific changes. Microsoft and Google declined to comment, and Meta and Apple had not responded to media requests. Reuters

At the same time, this warning comes amid broader tensions between state regulators and the federal government over AI regulation. The U.S. federal administration has been pushing legislation to limit or pre-empt state-level laws on AI, positioning this as a national competitiveness issue. State attorneys general argue that federal action shouldn’t prevent states from protecting their residents. Reuters

Why This Matters Beyond Tech Press Releases

This development is significant for several broader reasons:

✅ AI is no longer just a technological trend - it’s a social and legal concern affecting mental health, safety, and public policy.

✅ Regulators are shifting from dialogue to action, signaling that big tech can’t rely on goodwill alone when it comes to responsible AI deployment.

✅ Legal accountability is entering the AI conversation, expanding debates from ethics reports and corporate guidelines to real legislative and legal frameworks.

Looking Ahead: Regulation or Innovation Freeze?

Critics of strict regulation argue that AI innovation could be hampered if legal frameworks are too stringent or fragmented across states. Supporters of oversight counter that meaningful rules are essential to protect individuals - especially vulnerable populations - from harm caused by automated systems that are increasingly part of everyday life.

What’s clear is this: AI companies, policymakers, and users are entering a new phase where safety, legality, and societal impact are inseparable. The next steps -  how companies respond, how regulators enforce, and how the law evolves - will shape the AI landscape for years to come.