Head of Safety at OpenAI
A master student of artificial intelligence focused on exploring reinforcement learning and generative models with a background in software engineering.
How media typically covers Johannes Heidecke
Based on 6 scored articles
Research or work cited
OpenAI is rolling out GPT-5.1 with a warmer, more conversational tone and eight new customizable chat styles to all ChatGPT users.
“Referenced in "We're Rolling Out GPT-5.1 and New Customization Features. Ask Us Anything" in Reddit”
Referenced in coverage
“Left OpenAI following reorganization that integrated safety teams into research, one of several safety-focused leaders departing.”
“Johannes Heidecke, OpenAI's Head of Safety Systems, left the company in summer 2026.”
“Resigned sometime before July 10th, just before the Galaxy cyberattack incident.”
“Departing OpenAI as part of company reorganization that integrates safety work with model development.”
“Co-author on OpenAI's reinforcement learning research for beneficial model behavior.”