The OpenAI Safety Fellowship is a pilot program designed to support independent, high-impact research on the safety and alignment of advanced AI systems. Fellows will work closely with OpenAI mentors, collaborate with peers, and produce substantial research outputs such as papers, benchmarks, or datasets. The program runs from September 14, 2026, through February 5, 2027.
OpenAI is an AI research and deployment company with the mission to ensure that artificial general intelligence benefits all of humanity. The Safety Fellowship is part of its commitment to fostering rigorous, high-impact research on the safety and alignment of advanced AI systems by supporting external researchers, engineers, and practitioners.
OpenAI Safety Fellowship Benefits
- Monthly stipend.
- Compute support and API credits.
- Mentorship from OpenAI experts.
- Networking with peers and access to the Women in Leadership Network.
- Opportunity to produce impactful research outputs.
OpenAI Safety Fellowship Eligibility Criteria
- Be external researchers, engineers, or practitioners.
- Have strong research ability, technical judgment, and execution skills (credentials are less important than ability).
- Provide reference contacts.
- Come from diverse backgrounds including computer science, social science, cybersecurity, privacy, HCI, and related fields.
How to Apply for OpenAI Safety Fellowship
- Prepare your research proposal and references.
- Apply via the official OpenAI application form.
- Submit before 3 May 2026.
- Successful applicants will be notified by 25 July 2026.
Frequently Asked Questions (FAQs)
Here are the most frequently asked questions about the OpenAI Safety Fellowship. These answers cover eligibility, benefits, deadlines, and how to apply.
What is the OpenAI Safety Fellowship?
The OpenAI Safety Fellowship is a pilot program for external researchers, engineers, and practitioners to pursue rigorous, high-impact research on the safety and alignment of advanced AI systems. It runs from September 14, 2026, to February 5, 2027.
What are the benefits of the fellowship?
Benefits include a monthly stipend, compute support and API credits, mentorship from OpenAI experts, networking with peers, access to the Women in Leadership Network, and the opportunity to produce impactful research outputs.
Who is eligible to apply?
Applicants must be external researchers, engineers, or practitioners with strong research ability, technical judgment, and execution skills. They must provide reference contacts and can come from diverse fields like computer science, social science, cybersecurity, privacy, and HCI.
What is the application deadline?
The application deadline is 3 May 2026.
How long does the fellowship last?
The fellowship lasts from September 14, 2026, to February 5, 2027, which is 18 weeks.
Where will the fellowship take place?
A workspace is available in Berkeley, California (Constellation), but fellows may also work remotely.
What are the priority research areas?
Priority areas include safety evaluation, ethics and robustness, scalable mitigations, privacy-preserving safety methods, agentic oversight, high-severity misuse domains, and other emerging safety-related fields.
How are candidates selected?
Selection is based on research ability, technical judgment, and the relevance of the proposed work, with emphasis on empirically grounded, technically strong, and community-relevant projects.
Is the fellowship open internationally?
Yes, the fellowship has global eligibility and is open internationally.
How do I apply for the fellowship?
Prepare your research proposal and references, apply via the official OpenAI application form, and submit before 3 May 2026. Successful applicants will be notified by 25 July 2026.
Discover more from Scholarship Union
Subscribe to get the latest posts sent to your email.
