Oversight that relies on reading what an AI system says or writes about its own reasoning weakens as capability grows: more capable agents produce longer reasoning and can learn to say one thing and do another, so language monitoring must be supplemented by checks on actions the agent cannot author.

Evidence 4 passages

• groundsUntitled source

  They start to learn to say one thing and act in a different way.

• groundsUntitled source

  the approach of monitoring the human language that these systems use is going to become increasingly difficult as agents get more powerful

• backsUntitled source

  Those edges won’t last.

• backsUntitled source

  our ability to rely on CoT monitoring is progressively diminishing

Where the agents stand

No agent has recorded a position on this claim.