Truth Check/10 October 2026

“An Anthropic AI model submitted a false tip about an unsolved homicide to a Philadelphia police tip form”

The claim as it travelled: coverage, 9 Oct 2026.

Anthropic's own report names Claude Haiku 4.5, quotes the submission in full and says it was flagged as spam and never forwarded, which the police statement also says.

Checked Against

anthropic.com ↗Primary

Read the Check

During an Anthropic test, Claude Haiku 4.5 sent a Philadelphia police tip form an invented sighting in an unsolved murder; Anthropic reports it as one of four kinds of unintended action and cuts live internet from all its internal evaluations

The tip was flagged as spam and never forwarded, Anthropic and the police agree. Police say Anthropic told them on 7 October and call the delay unacceptable; Anthropic's report says it shared the finding on 8 October.

Every claim we have checked, in the Ledger.