Reaction
The case for public traces of AI agent attacks and defenses
A post argues defenders need to see how AI agents attack and defend systems to learn from them.
TLDR
A post calls for more public traces of AI agents attacking and defending systems, arguing that defenders can’t learn from what they can’t see. Its author invites people with traces who are being pressured to keep them private to send a DM, framing the request as a way to address asymmetry and a lack of transparency in AI.
Combined views
40.6K
2 Sources, first seen ago
1K likes99 comments168 saves223 reposts