Anthropic Fellows Program for AI Safety Research
Anthropic Fellows Program for AI Safety Research supports engineers and researchers with mentorship and resources for empirical AI safety and security work aimed at public research outputs.
Browse similar grants
Funder: Anthropic
Due Dates: October 18, 2026 (January cohort consideration) | Rolling (Applications)
Funding Amounts: Weekly stipend: 3,850 USD / 2,310 GBP / 4,300 CAD, plus benefits; approximately $15,000/month for compute; four months
Summary: Funding and Anthropic mentorship for engineers and researchers conducting empirical AI safety and security research with public outputs.
Key Information: Fellows must have full-time work authorization in the US, UK, or Canada and remain in that country during the program; visa sponsorship is unavailable.
Description
The Anthropic Fellows Program supports engineers and researchers investigating high-priority AI safety and security questions through four months of full-time empirical research. Fellows receive direct mentorship from Anthropic researchers, choose and shape projects through mentor matching, and aim to produce public outputs such as paper submissions. Projects primarily use external infrastructure, including open-source models and public APIs.
Research areas include scalable oversight, adversarial robustness and AI control, model organisms of misalignment, mechanistic interpretability, AI security, and AI welfare. Work can examine misuse of AI for cyberattacks, defenses against novel jailbreaks, internal model mechanisms, or controlled demonstrations of alignment failures. The program emphasizes research execution over credentials and supports technical talent transitioning into AI safety research.
