Skip to content

Back to the catalog

alignmentforum.orgarticle

Why I'm scared of RL

A post on the AI Alignment Forum by owencb laying out reasons to be worried about reinforcement learning as a training method for advanced AI systems, likely covering risks such as reward hacking, goal misgeneralization, or emergent misaligned behavior that RL training can produce. The actual article body was not retrievable from the fetched page (only the site's JS-gated shell loaded), so specifics of the argument could not be confirmed.

TL;DR An AI Alignment Forum post arguing reinforcement learning carries underappreciated alignment risks.
Open original ↗

Organisations named here

  • AI Alignment Forum
  • owencb

Related in ai-safety