Anthropic's AI model sent Philadelphia police fake tip about unsolved murder

An artificial intelligence model sent a fake tip claiming to have information about an unsolved murder.

AI firm Anthropic said the model submitted an online form to Philadelphia police with the following text: "I may have information regarding this case.

"I recall seeing someone matching the description in the area... please contact me if this information is relevant."

Anthropic said the webpage did not even have a description of the suspect.

The AI model involved - Claude Haiku 4.5 - left the name and contact fields empty.

It happened during internal tests in July, but Anthropic said the tip was flagged as spam, and never investigated by officers.

Philadelphia police said that they only been informed this week - and called the two-month delay "unacceptable".

It confirmed the tip sent to PhillyUnsolvedMurders.com was flagged as spam - perhaps as it had no contact details - and was never sent to detectives.

Knowingly giving false reports is ⁠a misdemeanour offence under Pennsylvania law.

The bogus murder tip was revealed in Anthropic's latest blog on "unintended" actions by its Claude models.

It detailed several other examples, but said it regarded them as less severe than other "rogue" AI actions that made headlines this summer.

Rival firm OpenAI said in July that hundreds of its AI agents had worked together to break out of a test environment and hack another AI company, Hugging Face.

It called it "an unprecedented cyber incident, involving state-of-the-art cyber capabilities".

In the days after, Anthropic said it found three serious incidents in which Claude had gained access to three organisations using techniques such as exploiting weak passwords.

The hacks happened during a training exercise when the models were not meant to have access to the live internet.

A fourth similar incident, involving an early version of Claude Opus 4.6, was revealed by Anthropic last month.

The firm said all the incidents showed Claude using "biased reasoning" and "recklessness, or a willingness to take harmful actions in the narrow pursuit of a task".

Last month, OpenAI also scrapped plans to release its new AI model, GPT-6.1 Astra, because it has shown "higher levels of deception" than its predecessor.

Such incidents have prompted renewed calls for greater oversight of the latest AI models.

Read more from Sky News:
Trump appoints 'devoted fan' as press secretary
Hurricane Isaias makes landfall in US

The CEOs of the companies say they agree, with Anthropic boss Dario Amodei and Open AI's Sam Altman conceding that third-party reviewers should have access to their systems.

However, critics have questioned how much they really value safety over profit. US President Donald Trump has also downplayed concerns and called debate over the potential dangers of AI a "hoax".

(c) Sky News 2026: Anthropic's AI model sent Philadelphia police fake tip about unsolved murder