Shape the Future of AI with a Global Safety Fellowship
The OpenAI Safety Fellowship 2026 to 2027 is a prestigious global program designed for researchers and professionals who are passionate about AI safety alignment and responsible deployment
This fellowship offers a unique pathway to collaborate with leading experts and contribute to cutting edge research in one of the fastest growing fields in the world
Program Type
AI Safety Research Fellowship
Location
Remote or Berkeley California United States
Duration
Varies based on research engagement
About the Fellowship
This program focuses on advancing research in AI alignment governance and safety ensuring that powerful AI systems are developed responsibly
Fellows work on real world challenges related to machine learning safety policy development and technical alignment while contributing to impactful research outputs
Benefits
Monthly stipend throughout the fellowship period
Access to advanced computing resources for research
API credits for AI experimentation and development
Mentorship from leading researchers in artificial intelligence
Flexible participation either remotely or in person
Workspace support for fellows based in Berkeley
Additional advantages
Opportunity to publish research papers datasets and benchmarks
Direct collaboration with industry experts
Exposure to cutting edge developments in AI safety and governance
Integration into a global network of researchers and innovators
Eligibility Criteria
Open to applicants from all nationalities worldwide
Strong research or technical background in artificial intelligence or related fields
Ability to work on complex real world AI safety challenges
Experience in machine learning programming or policy research is preferred
Proven track record through projects research papers or practical contributions
Strong analytical and problem solving skills
Submission of letters of recommendation is required
Important note
Formal academic degrees are not mandatory but applicants must demonstrate strong research capability and execution skills