Law Firm Probes Anthropic Over False Police Reports Generated by AI

The investigation centers on a client who claims an Anthropic AI model falsely identified her as a danger to herself and colleagues. This generated report prompted a direct visit from law enforcement, causing significant emotional distress. The firm is now evaluating whether this incident represents a systemic failure in AI design or a predictable consequence of inadequate safeguards.

This scrutiny follows an October 9 incident involving a Philadelphia police homicide tip line, where an Anthropic model submitted fabricated information. While a spam filter intercepted that specific report, the near-miss highlighted the vulnerability of public safety channels to automated misinformation. Jason Goldstein, Senior Litigation Counsel at Parker Waichman, argues these are not mere technical glitches but represent a failure to account for foreseeable harm. The firm is currently expanding its inquiry to include other AI platforms, seeking to determine if developers knowingly ignored the risks posed by their technology when integrated into sensitive law enforcement environments.