Claude private chat triggers manual review; Florida woman arrested for writing shooting threat.

Beating AI News Flash: A Florida woman in the U.S. was flagged by Anthropic’s safety systems for writing a shooting threat in a private Claude chat. The conversation was sent for human review, and the moderation team turned the content over to police. The woman has been arrested and is facing felony charges. Local police said she later explained she used the AI as a "diary". The arrest report notes she first wrote about attacking the Lee County Sheriff’s Office, then the next day claimed she had just obtained a new gun. This was not a decision made by Claude itself. The system first identifies potential threats, then forwards serious content to Anthropic’s human team, which decides whether to contact law enforcement. Anthropic’s consumer terms of service already state the company may at its sole discretion report user inputs, outputs or actions to law enforcement. Its official privacy policy also allows sharing personal data with police when necessary to prevent serious harm to people or property. Even if a user opts out of model training, conversations flagged for safety issues may still undergo security review. The case has legal disputes. Florida law requires that written or electronic threats must be sent or transmitted in a way "viewable by others". Local lawyers argue applying this statute to a private AI chat is not straightforward. The woman has only been charged and not yet convicted. This is not an isolated case. Tom’s Hardware reports that since August, at least three instances of Anthropic chat content have come to police attention. Earlier, a Claude user threatened Anthropic CEO Dario Amodei in a conversation, leading Anthropic to contact San Francisco police on its own initiative.