Anthropic Fellows Program, AI Security
About This Role
This is a fully-funded, four-month fellowship offering direct mentorship from Anthropic researchers, access to shared workspaces in Berkeley or London, and generous compute funding. Much like early contributors to pioneering blockchain networks, Fellows get to help define the security standards of an emerging and consequential technology—AI—before the frameworks are set in stone.
What You'll Be Doing
Conduct empirical research into AI security vulnerabilities, drawing parallels to smart contract auditing and decentralized system hardening
Collaborate with Anthropic's research team to publish findings, with over 80% of past Fellows producing public papers
Use open-source models and public APIs to test and stress security assumptions in deployed AI systems
Explore attack surfaces and defense mechanisms relevant to safe, distributed AI deployment
Engage with the broader AI safety and security research community, similar to how web3 builders engage open-source communities
Translate experimental results into actionable security recommendations
What We're Looking For
Strong technical foundation in computer science, mathematics, or physics
Genuine motivation to ensure AI—like blockchain and decentralized systems—is built safely and responsibly for society
Comfort working in fast-paced, collaborative, and ambiguous research environments
Interest in transitioning into empirical AI research, potentially full-time at Anthropic
Ability to work independently with external infrastructure and resources
Strong communication skills for producing public-facing research output
Bonus Points
Prior experience in security research, including crypto, blockchain, or distributed systems security
Familiarity with adversarial machine learning or red-teaming methodologies
Experience contributing to open-source research or technical communities
Background in applied cryptography or decentralized protocol design
Why Join Us
Weekly stipend of $3,850 USD / £2,310 GBP / $4,300 CAD plus benefits
~$15k/month in compute and research funding
Direct mentorship from leading AI safety and security researchers
Opportunity to influence the security foundations of transformative AI—comparable in significance to securing the future of decentralized finance and web3 infrastructure
Ready to Apply?
To submit your application for this position, please visit Anthropic's official website and follow their application process.