Re:Investigating three real-world incidents in our cybersecu...
by willy_k on Friday July 31, 2026 @01:17PM
(#49125996)
Yet, the models are objects built by humans. Any and all responsibility inevitably falls on the humans that built something with those flaws, and/or allowed them to exercise said flaws.
Re:Investigating three real-world incidents in our cybersecu...
by bayindirh on Friday July 31, 2026 @03:47PM
(#49127787)
In short, it’s either a “not implemented” case or a bug.
There’s no other explanation. Complexity of the algorithm or the application doesn’t change the reality.
Re:Investigating three real-world incidents in our cybersecu...
by matheusmoreira on Friday July 31, 2026 @06:26PM
(#49129234)
> some runs “rationalized that the real company must be part of the exercise”
How human of them.
If the model saw signs, but “subconsciously” (below the level of reasoning traces) chose to turn a blind eye to them, out of a relentless focus on achieving the objective, then that absolutely is the model’s “fault”, i.e. a case of misalignment of the sort which will become increasingly dangerous over time.
The blog post mentions that some runs “rationalized that the real company must be part of the exercise” and to me that seems suspiciously like the latter.
Hacking can be patched with classifiers and with better sandboxes, but this is a much more general problem. Fundamentally, we need be able to trust that models will be honest with users and with themselves. This applies at some level to almost every LLM interaction.