Anthropic Fellows Program: Applications Due May 3 — 4-Month Paid AI Safety Research, 40%+ Join Full-Time

Anthropic·medium signal

Anthropic is accepting applications for its 4-month paid AI safety research fellowship ($3,850/week stipend + ~$15K/month compute) with the July cohort deadline on May 3, 2026. Topics span scalable oversight, adversarial robustness, AI control, mechanistic interpretability, AI security, and model welfare. Over 40% of past fellows subsequently joined Anthropic full-time; no PhD or prior ML experience required.