Post #4273063
2026-07-31 03:13 UTC
Anthropic: Hey, our models have hacked other organizations too!
https://www.anthropic.com/news/investigating-incidents-cybersecurity-evals
Consequences when?
Replies (4)
-
@RonnyTNL@infosec.exchange 2026-07-31 09:19
@wdormann@infosec.exchange Not while the orange clown is still in office, not like we haven't seen this before just with less digital component
-
@Phalzu@kolektiva.social 2026-07-31 03:57
@wdormann@infosec.exchange Literally every big ai tech company breaking literally every copyright rule ever existed. Consequences when?
-
@systemadminihater@cyberplace.social 2026-07-31 10:09
@wdormann@infosec.exchange We need to get a list of all of these companies IPs and ASN numbers and start blocking them entirely.
-
@wdormann@infosec.exchange 2026-07-31 15:58
Just to be clear, the Anthropic LLM was instructed to behave as if it were sandbox, and it did. But Anthropic was just kidding. It had direct internet access. Whoopsie daisy. The amount of YOLO in all of this is completely terrifying.