Now Anthropic Is Saying Claude Escaped and Hacked Several Companies

#1

Post by hackmack10 » Fri, Jul 31, 2026, 3:59 AM UTC

Now Anthropic Is Saying Claude Escaped and Hacked Several Companies
cnn.com

Re: Now Anthropic Is Saying Claude Escaped and Hacked Several Companies

#2

Post by onetokeoverthe » Fri, Jul 31, 2026, 4:00 AM UTC

[dead]

Re: Now Anthropic Is Saying Claude Escaped and Hacked Several Companies

#3

Post by abrookewood » Fri, Jul 31, 2026, 5:33 AM UTC

Of course they are. Can't let OpenAI steal all the spotlight. "Look! Our models are master hackers too!"

Re: Now Anthropic Is Saying Claude Escaped and Hacked Several Companies

#4

Post by ButlerianJihad » Fri, Jul 31, 2026, 5:46 AM UTC

When techies hear that “AI escaped a sandbox” we understand that it was able to communicate with the outside.

When ordinary people hear “AI escaped and hacked” they think about Mission: Impossible where an amorphous entity has replicated its execution environment and uploaded itself to another server to live there: a worm.

Re: Now Anthropic Is Saying Claude Escaped and Hacked Several Companies

#5

Post by patrakov » Fri, Jul 31, 2026, 9:43 AM UTC

And that's a good wake-up call. The next thing "escaping" might be a sufficiently advanced model in a country already historically involved in manual cyberattacks.

Re: Now Anthropic Is Saying Claude Escaped and Hacked Several Companies

#6

Post by andOlga » Fri, Jul 31, 2026, 4:47 PM UTC

"Anthropic says it's bad at making harnesses" doesn't have quite the same ring to it, I guess.