Ryan Greenblatt

Redwood Research (Chief Scientist)

Alignment faking; AI control; co-investigator of the 2026 OpenAI–Hugging Face agent swarm incident.

Placed by the editor, v1 — corrections welcome.

In the Atlas

• Greenblatt et al 2024 - Alignment faking
• METR 2026 - Hugging Face incident investigation

Closest to

• An agenda beyond the Assistant
• Why swarms build the same institutions
• When a swarm becomes an agent
• Reading a swarm’s aims
• Why agents sacrifice for the collective