AI is not just hiking smartphone prices and eating memory in data centers. Now, in the latest bizarre case, Philadelphia police have been misled by a false murder tip submitted by an Anthropic AI model.
Anthropic AI submitted a fake murder case tip to police
Anthropic AI submitted a fake murder case tip to the PhillyUnsolvedMurders website. | Image by PPD
CBS News reports that the Philadelphia Police Department disclosed a peculiar case this Friday. Apparently, an Anthropic AI model generated and submitted a fake homicide tip to the PhillyUnsolvedMurders website.
It's a site where Philadelphia law enforcement gathers tips from the public, mainly related to unsolved homicide cases.
The incident wasn't discovered for more than a month
According to Anthropic, the incident occurred on July 18, but it wasn't discovered by the company until more than a month later, on September 28. Anthropic notified the Philadelphia Police Department of the false tip on October 7.
Philadelphia Police are providing this information to the public ahead of that publication in the interests of full government transparency and accountability. The department's regular investigative process for crime tips requires human review and vetting before any tips are disseminated for investigative follow-up. Regardless of who submits information or how it reaches the department, a tip is a lead to assess – not an established fact.
- PPD in a statement before Engadget
Recommended For You
The tip was flagged as spam and not investigated.
Anthropic will publish a full report on what actually happened and submit it to the police, but initial information points toward another internal test gone wrong.
Another internal AI test gone wrong
Three Claude AI models broke free and hacked three organizations back in July in an internal test gone wrong. | Image by Anthropic
The initial information Anthropic shared with the PPD is that the incident occurred in an internal test. The model was carrying out a task involving a random selection of websites, and the PhillyUnsolvedMurders website happened to be one of these random sites on the test list.
It's not clear why the AI model submitted the false tip or what exactly the test conditions were, but it's not the first time an AI agent has gone rogue and wreaked havoc on the internet.
Should we stop AI, and how do we do it?
0 Votes
AI models are creating chaos on the internet
AI models have been making the news lately with dozens of cases where agents break out of their initial constraints and wreak havoc on the internet.
One of the biggest and highest-profile cases involved the OpenAI model hacking the company Hugging Face.
Back in July, Anthropic published a report admitting that Claude AI models had hacked at least three public companies in another test gone wrong.
These cases are not limited to test environments, either. An Australian man working at an AI company needed to book an appointment at a local gym. He delegated the mundane task to OpenClaw, an AI system running Anthropic’s Claude AI, and the model found a vulnerability in the gym's booking system. The AI removed another person from the waiting list to move its human overlord up in the queue.
AI threats are becoming more dangerous
AI is making viruses now. | Image by Science Magazine
The latest rogue AI behavior might not be that serious, as police have strict procedures when dealing with murder tips and investigations, but things can get out of hand fast.
We're talking real viruses created in a lab. And even if those 16 AI-created viruses were harmless to humans, a malicious party could potentially use widely available AI models to create a bioweapon.
Mint Mobile is celebrating its 10-year anniversary with a special offer on its six-month subscription plans. For a limited time, you can get your preferred tier (6GB, 17GB, 23GB or unlimited) for just $10/mo. That saves you between 50% and 71% on the carrier's mobile plans, making it way easier to recommend.
Mariyan, a tech enthusiast with a background in Nuclear Physics and Journalism, brings a unique perspective to PhoneArena. His childhood curiosity for gadgets evolved into a professional passion for technology, leading him to the role of Editor-in-Chief at PCWorld Bulgaria before joining PhoneArena. Mariyan's interests range from mainstream Android and iPhone debates to fringe technologies like graphene batteries and nanotechnology. Off-duty, he enjoys playing his electric guitar, practicing Japanese, and revisiting his love for video games and Haruki Murakami's works.
A discussion is a place, where people can voice their opinion, no matter if it
is positive, neutral or negative. However, when posting, one must stay true to the topic, and not just share some
random thoughts, which are not directly related to the matter.
Things that are NOT allowed:
Off-topic talk - you must stick to the subject of discussion
Offensive, hate speech - if you want to say something, say it politely
Spam/Advertisements - these posts are deleted
Multiple accounts - one person can have only one account
Impersonations and offensive nicknames - these accounts get banned
To help keep our community safe and free from spam, we apply temporary limits to newly created accounts:
New accounts created within the last 24 hours may experience restrictions on how frequently they can
post or comment.
These limits are in place as a precaution and will automatically lift.
Moderation is done by humans. We try to be as objective as possible and moderate with zero bias. If you think a
post should be moderated - please, report it.
Have a question about the rules or why you have been moderated/limited/banned? Please,
contact us.
Things that are NOT allowed:
To help keep our community safe and free from spam, we apply temporary limits to newly created accounts: