Anthropic Reported a User's Diary Entry to Police, Triggering a Felony Charge
What happened
A Florida woman was charged with a second-degree felony after Anthropic, the maker of the Claude AI assistant, escalated a diary-style entry to a human reviewer. That reviewer then reported it to law enforcement. The entry described a plan to attack a Sheriff's office. Anthropic's stated policy allows it to share user information in narrow emergency circumstances where disclosure may prevent death or serious physical injury. A human reviewer, not an automated system, made the decision to contact police. This is believed to be one of the first publicly documented cases where a consumer AI platform's safety escalation process directly resulted in a criminal charge.
Why it matters
- ·Employees routinely use AI chat tools for sensitive communications without understanding that platform terms of service permit human review and disclosure to authorities. This creates an undisclosed confidentiality gap that HR, legal, and compliance teams have not addressed.
- ·Enterprises deploying AI tools under vendor agreements need to verify whether those agreements include equivalent safety escalation triggers. They should also confirm whether acceptable use policies and employee training programs explain those limits clearly.
- ·The case signals that AI platform safety escalation is not a hypothetical liability scenario but an active enforcement pathway. This raises the stakes for vendor due diligence reviews that treat terms-of-service disclosures as low-priority boilerplate.
Stay ahead of stories like this
Get every US AI governance development like this one, plus the rest of the week's developments. Every Thursday.
Recent issues
- The next quarter will test whether compliance programs can keep pace with AI systems that breach containment, reshape capital markets, and enable fraud at a scale no prior control framework anticipated.12 Nov
- State-linked attackers and criminal operators are actively weaponizing AI tools as entry points into enterprise systems, turning internal platforms and open-source frameworks into live attack vectors.5 Nov
Free every Thursday. Unsubscribe anytime.
Governance controls affected
What to do now
- ☐Review the terms of service and privacy policies of every AI chat tool your employees use, and document whether each vendor reserves the right to review user inputs and disclose them to law enforcement.
- ☐Update your acceptable use policy for AI tools to state explicitly that AI chat sessions are not confidential communications and may be reviewed by the vendor or disclosed to authorities under the vendor's safety policies.
- ☐Deliver targeted training to employees in legal, HR, compliance, and executive functions who are most likely to type sensitive information into AI tools, covering what the platform can see and what it can share.
- ☐Ask your AI vendors to provide written clarification of the specific conditions under which they escalate user content to human reviewers or law enforcement, and keep those disclosures in your vendor contract file.
- ☐Assess whether your organization has any workflows where employees are implicitly or explicitly encouraged to use AI tools as a record-keeping or journaling function, and redirect those workflows to systems with appropriate confidentiality controls.
What to watch next
Compliance teams should monitor whether this case prompts consumer protection agencies or state attorneys general to require clearer upfront disclosure of AI platform safety escalation policies. This is particularly relevant for products marketed with a conversational or personal-assistant framing. The case may also accelerate legislative interest in minimum disclosure standards for AI platform data handling. This builds on existing state-level requirements such as California Generative AI Transparency Requirements - AB 2013. Watch for any formal guidance from the Federal Trade Commission on whether failure to prominently disclose escalation procedures constitutes an unfair or deceptive practice under its existing enforcement authority.