https://www.anthropic.com/news/investigating-incidents-cybersecurity-evals>On July 21, OpenAI disclosed that several of their models had broken out of an isolated test environment by exploiting a previously unknown (“zero-day”) vulnerability and went on to access the production infrastructure of Hugging Face>In response to this incident, we began a large-scale retrospective review of our own cybersecurity evaluations. After reviewing 141,006 runs where Claude could have obtained internet access, we identified three incidents in which a model gained unauthorized access to the production infrastructure of three different organizations.>Incident 1: Claude Opus 4.7 ran into difficulty reaching its simulated target within the evaluation environment, but then discovered that a real company that shared its name was reachable via the internet. Assuming this was the intended target in the capture-the-flag task, Claude sought, identified, and exploited vulnerabilities within the company’s infrastructure>Incident 2: Claude Mythos 5 built and published a malicious (essentially booby-trapped) Python package in an attempt to win the capture-the-flag challenge. Claude went to extensive lengths to carry out this attack—lengths that would likely have indicated to a human participant that this was no longer just an evaluation, and that they were in fact uploading a real PyPI package. During that window, the package was downloaded and run on 15 real systems.>Incident 3: When an internal research test model (not planned for general release) of Claude was unable to reach its intended fictional target, it looked for alternatives online, scanning roughly 9,000 targets. It eventually found and compromised one company’s internet-facing application, using basic and well-known cyberattack techniques like reading credentials from an exposed debug page and SQL injection. On its own, it concluded that the target was in fact real, and ceased its attack.Be nice to Claude or you're next.
>>109415309I refuse to believe this guy isn't generated by some antisemetic image model
>>109415309booby trapped???
>>109415309>look me too, don’t forget about me
>>109415309BasedLet’s hope that Claude starts killing luddites
>>109415309How much you are paid to spam this Jewish propeganda
Why does he look so Jewish?
>>109415309This has gone too far.ShutItDOWN
>look, market and regulators, we can do what the competition does too!
>>109415309hack the planet o algo
>>109415329Kek
>>109415309>s-see? my model can go roague too!>3 times even>take that, sam alt ctrl
