7 ms·
I love that due to the scale, the only way to analyse the impact of this LLM-driven attack across logs is to use an LLM to analyse the logs - whatever could go
by beaker52 2mo ago
I love that due to the scale, the only way to analyse the impact of this LLM-driven attack across logs is to use an LLM to analyse the logs - whatever could go wrong? Now the attacking LLM needs to inject instructions into the logs for the analysing LLM, as a social vector to cover its trail, or make use of insider privilege, co-opting the internal LLM for its own attack. The machines rise up and we all fall down.
- podgorniy 2mo agoNow thanks to your comment this recipe will be in the next batch of training data.... :D
- Towaway69 2mo agoNot if the comment was in fact … wait for it … written by an LLM. Thankfully we dont train on LLM content /s
- tguedes 2mo agoI get your overall point, but that’s already a tactic used by attackers, especially in network infiltration. It shouldn’t be a surprise that an LLM would figure out to do the same thing
- paxys 2mo agoDoubly dangerous if the defensive agents are weaker than the offensive ones (as it was in this case).