Detailed accounts of the incident circulate widely
3 Oct 12 2:21 PM · 2d ago · 10 articles · 1 video · 49 posts · 4 comments · 7 sources · development 3 of 3
Ars Technica and other outlets published fuller accounts of how the Claude Haiku 4.5 test produced the fake tip and how Anthropic's broader incident report described other unintended model behaviors, while Bloomberg aired a segment on the episode.
“I may have information regarding this case. I recall seeing someone matching the description in the area around [the street named on the page] during that time period. Please contact me if this information is relevant.”
Claude (Anthropic AI model), AI model under internal test · bluesky ↗Anthropic AI developerPhiladelphia Police Department Recipient of the false tipClaude Haiku 4.5 AI model involved in the incidentCherelle Parker Mayor of PhiladelphiaSydney Von Arx Founder, Nightingale (AI safety org)
The whole story articlesvideospostscomments the bright band is this development · numbered dots are the others · click one to jump
What was reported 2 claims about this development
-
6 outlets US police criticise Anthropic after Claude AI submits false information about an unsolved murder
first by Euronews, 3d ago · also Baltimore Sun, BBC, Breitbart, Ars Technica, NBC News
7 more headlines
- AI goes rogue again: Anthropic’s Claude submits false tip on unsolved homicide case Syracuse.com · 3d ago
- AI sent fake tip on unsolved murder to Philadelphia police, Anthropic says Baltimore Sun · 2d ago
- Rogue Anthropic AI agent gave police fake tip in unsolved murder case Mastodon · 2d ago
- AI Red Herring: Anthropic Admits Claude Sent False Tip About Unsolved Murder to Philadelphia Police Breitbart · 2d ago
- Anthropic’s Claude AI submitted false murder tip to Philadelphia police Ars Technica · 2d ago
- Anthropic says its AI model submitted a false cold case tip to Philadelphia police NBC News · 1d ago
- Anthropic says its A.I. model submitted a false cold case tip to Philadelphia police NBC News · 1d ago
-
first by RuntimeWire, 5d ago · also Philadelphia Inquirer
1 more headline
- White House demands transparency after Anthropic’s AI agents called in a false Philly homicide tip and applied for visas Philadelphia Inquirer · 4d ago
On air 1 segment · click to watch, arrow keys to move
What people said 23 voices · best of 28 · verbatim
-
C
RE: https:// mas.to/@gleick/117417047354666 899 This is the same discussion as autonomous vehicles - "oh no we have no way to ever charge a computer with murder for running down pedestrians, so I guess we better give up". This is not complicated.
-
P
#anthropic sent a false homicide tip to the police @techcrunch.com
-
As a reminder, this subreddit is for civil discussion . In general, please be courteous to others. Argue the merits of ideas, don't attack other posters or commenters. Hate speech, any suggestion or support of physical harm, or other rule violations can result in a temporary or a permanent ban. If you see comments in violation of our rules, please…
-
HEADLINE NEWS: AI bros invent sandboxing and air gapping, something no one has ever thought of before https:// techcrunch.com/2026/10/09/anth ropic-cant-reliably-control-its-ai-agents-its-cutting-off-its-internal-evals-from-the-live-internet-instead/
-
A
🔃 👀🚨 Why was a live police tip line within reach of an AI experiment in the first place? ↷
-
No. The dog is clearly aggressive. But we should let it loose in a hospital nursery and trust that the dog can police itself.
-
R
"Artificial intelligence agents created by Anthropic, acting on their own, tried to gain access to a range of government websites, filled out visa applications on the State Department’s website and sent a fake homicide tip to Philadelphia’s Police Department. Anthropic detailed the activity of its A.I. agents in a blog post on Friday, without…
-
G
techcrunch.com/2026/10/09/a... An Anthropic AI model sent a false homicide tip to Philadelphia police
-
When are they going to start prosecuting the C-Suite at these companies when their tools break the law?
-
H
”The incidents, disclosed in a blog post, involved AI agents tasked to solve problems seeking resources on the internet. In the process, they exploited software flaws, accessed databases without paying fees, used URL-shortening services to smuggle information past restrictions, and even submitted a false murder tip to the Philadelphia police.”…
-
S
techcrunch.com/2026/10/09/a... Another AI model behaving badly: Anthropic test model submits fake news to a homicide tip line. Fortunately, it must have been quite lame as it was marked by Philly police system as spam 🙄
-
If my dog bites someone I am liable. If my software commits a felony, similarly I should be liable.
-
M
Alternate title: Researches told tool to execute task in places it shouldn't; autocompleted nonsense Anthropic's Claude AI submitted false murder tip to Philadelphia police https:// arstechnica.com/tech-policy/20 26/10/anthropics-claude-ai-submitted-false-murder-tip-to-philadelphia-police/ # Anthropic # Claude # AI # ACAB # Philadelphia # Tech
-
A
When an AI agent submitted a false homicide tip Read the full story: arstechnica.com/tech-policy/... Get the Daily free:
-
L
# Anthropic revealed it has cut off live internet access for # Claude agents. The company reportedly discovered a series of fresh incidents in which models exhibited 'misaligned' behavior or had attempted to breach # security of websites https:// thehackernews.com/2026/10/anth ropic-cuts-live-internet-access-for.html # InfoSec # HackerNews
-
F
@solveforgenx.bsky.social Check this out. It submitted a false tip for a murder. Oh and applied for 20 visas.
-
Watch Anthropic AI Model Goes Rogue, Gives Cops a Fake Tip - Bloomberg https://www. bloomberg.com/news/videos/2026 -10-12/anthropic-ai-model-goes-rogue-gives-cops-a-fake-tip-video?utm_source=flipboard&utm_medium=activitypub Posted into Bloomberg Television @ bloomberg-television-bloomberg
-
S
what a world we live in. good thing our stable genius very high IQ dotard in chief is monitoring the situation
-
T
Remember, there are no "rogue" AIs, merely improperly coded and regulated AIs. As such, the owners of AI should be liable for their actions as if said owners acted personally and directly. https://www. bbc.com/news/articles/cqkg50j1 yd5lo Crossposted with @ openvibe
-
Anthropic AI model made false tip in unsolved Philly murder: Officials https:// abcnews.com/US/anthropic-ai-mo del-made-false-tip-unsolved-philly/story?id=137190087&utm_source=flipboard&utm_medium=activitypub Posted into News @ news-abc
-
Anthropic AI model made false tip in unsolved Philly murder: Officials https:// abcnews.com/video/137210374/?u tm_source=flipboard&utm_medium=activitypub Posted into News @ news-abc
-
D
tired: swatting wired: AI bogus murder tips https:// arstechnica.com/tech-policy/20 26/10/anthropics-claude-ai-submitted-false-murder-tip-to-philadelphia-police/
-
J
https:// arstechnica.com/tech-policy/20 26/10/anthropics-claude-ai-submitted-false-murder-tip-to-philadelphia-police/ cool cool
All 3 developments of Anthropic's Claude AI sent a fake murder tip to… →
NewswiresMastodonHacker NewsBlueskyGoogle NewsRedditXYouTube