OpenAI Safety Fellowship 2026–2027 in USA (Fully Funded)
Applications for the OpenAI Safety Fellowship 2026–2027 are now available. The OpenAI fellowship is a new research program in the United States that gives selected people money and support to work on important AI safety and alignment problems. It runs from September 14 2026 to February 5 2027. You get a monthly stipend compute resources API credits and mentorship from OpenAI experts. You can work remotely or in person in Berkeley. The program wants real research outputs like papers, datasets or benchmarks. Also see UNESCO MGIEP Youth Fellowship 2026 | Fully Funded South korea.
OpenAI Safety Fellowship Overview
| Country | United States |
| Program | OpenAI Safety Fellowship |
| Duration | About 5 months |
| Coverage | Fully Funded |
| Eligible For | Researchers, engineers, and practitioners interested in AI safety |
| Application Deadline | May 3, 2026 |
What is the OpenAI Safety Fellowship 2026–2027?
OpenAI started this new fellowship to support outside researchers who want to study AI safety and alignment. As AI systems become more powerful it is very important to make sure they are safe and helpful. You work on real problems like misuse, bias, privacy, robustness and control of advanced AI. The program is short and focused so you produce useful work by the end. It is good for people who already have some research experience and want to help make AI safer for everyone.
Fellowship Benefits
You get good support during the program.
- Monthly stipend to cover living costs.
- Dedicated compute resources for your research.
- API credits to test AI models.
- Mentorship from OpenAI safety researchers.
- Workspace in Berkeley if you join in person (through Constellation).
- Chance to work remotely if you prefer.
- Opportunity to create papers datasets or benchmarks.
Also See: DAAD JSPS Short-Term Research Fellowship 2026 in Japan
Eligibility Criteria
- You can be from any country.
- You should have strong research or technical skills.
- You need experience in computer science, cybersecurity, social sciences, human-computer interaction or related fields.
- You must show you can do serious work on AI safety questions.
- You need good letters of recommendation.
- The program is open to people with different backgrounds not just those from top universities.
Required Documents
You need these papers.
- Completed application form.
- CV or resume with your experience.
- Strong motivation statement about your interest in AI safety.
- Research proposal or idea.
- Letters of recommendation.
- Any previous work samples or publications if you have them.
How to Apply for OpenAI Safety Fellowship
- Go to the official OpenAI Safety Fellowship page.
- Read all details and requirements carefully.
- Prepare your motivation statement and research ideas.
- Fill the online application form.
- Upload the documents mentioned above.
- Submit before the deadline.
- They review all applications.
- Shortlisted people get notified by July 25 2026.
Fellowship Dates / Deadline
- Applications are open now.
- Deadline is May 3 2026.
- Results come by July 25 2026.
- Program runs from 14 September 2026 to 5 February 2027.




