Tag: model-safety

All the articles with the tag "model-safety".

• OpenAI Paused RL on Its Latest Models. Safety Is Real—and So Is the Business Logic.

  21 Aug, 2026

  OpenAI’s two-week RL pause reflects a genuine cyber risk, but it also sends a commercial signal to regulators and investors.