OpenAI· GPT / ChatGPT / APICapability change

From hard refusals to safe-completions: toward output-centric safety training

OpenAI describes a new safe-completions approach in GPT-5 that replaces hard refusals with output-centric safety training, aiming to improve both safety and helpfulness on dual-use prompts.

From the source

Discover how OpenAI's new safe-completions approach in GPT-5 improves both safety and helpfulness in AI responses—moving beyond hard refusals to nuanced, output-centric safety training for handling dual-use prompts.

openai.com

Announced

07 Aug 2025 00:00 UTC

Detected

03 Aug 2026 21:11 UTC (+361d after)