AI security and the argument for a slowdown
A user argues that security progress hasn't kept pace with AI, calling that gap perhaps the strongest argument for a slowdown.
TLDR
A user argues that AI security progress hasn't matched the pace of AI and suggests this may be the strongest argument for a slowdown. They distinguish security itself from safety rebranded as security to avoid the term “safety.” They also recommend a RAND report on protecting algorithmic secrets.
Combined views
701
1 Source, first seen ago