Johannes Heidecke, OpenAI's Head of Safety Systems, has told colleagues he is leaving, and the company is folding its safety team into its research organization. The move was reported by WIRED based on an internal memo circulated around July 10, 2026 by Chief Research Officer Mark Chen.
July 10, 2026 · OpenAI
Safety Chief Departs as OpenAI Folds Safety Into Research
Head of Safety Systems Johannes Heidecke is leaving. In an internal memo, Chief Research Officer Mark Chen said the safety team will be absorbed into the research organization — as training accelerates and release cycles shorten, coordination demands have grown larger than ever.
2021
Heidecke joined as an AI Safety Analyst; became Head of Safety Systems in 2024.
4+
Senior safety-linked exits named across the reorganization in recent years.
GPT-5.6
Latest fast-shipped model — its own system card noted some misaligned behavior.
New reporting line: safety under research
Departing
Johannes Heidecke
Head of Safety Systems
→
Interim head
Saachi Jain
Safety Systems, reports to Glaese
→
Now oversees safety
Mia Glaese
VP of Research and Safety
The company's case
Embed safety review early and directly in model, product and release decisions.
Deployment Safety Hub launched to search and view system cards.
Fits industry trend — Anthropic and Google DeepMind also consolidating safety teams.
The concern
Continues a string of safety-researcher exits read as talent outflow.
Follows the earlier dissolution of the Superalignment team.
Heidecke had warned next-gen models could reach a "High" bioweapons-risk classification.
The shared industry question:
how to embed the balance of frontier-model speed and safety into an organization — as release cycles shorten and coordination challenges grow.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…