Singapore AI Safety Fellowship opens applications for paid research program

Selected researchers will receive SGD 5,000 a month, housing, travel, mentorship, and up to USD 30,000 in compute during the three-month Singapore cohort.

Applications for the Singapore AI Safety Fellowship close on July 10, with the residential program running from September 21 to December 4, 2026

Applications have opened for the Singapore AI Safety Fellowship, a three-month residential research program focused on technical AI safety, governance, and international research collaboration.

The full-time fellowship will run in Singapore from September 21 to December 4, 2026. Applications close on July 10.

Participants will receive a monthly stipend of SGD 5,000, with housing and travel to and from Singapore covered. Individual research projects can also receive up to USD 30,000 in computing resources.

The program is open to emerging researchers with a record of technical research, an interest in AI safety, and the ability to work across cultures and academic disciplines. Successful applicants will relocate to Singapore for the full program, with visas arranged by the fellowship.

SASH, which is running the program, has confirmed mentors from universities and AI organizations including the National University of Singapore, Tsinghua University, Fudan University, Anthropic, Concordia AI, FAR.AI, and the Oxford Martin AI Governance Initiative.

Applications close on July 10

Applicants will begin with a written application that SASH estimates will take approximately one hour to complete.

Shortlisted candidates will then complete standardized assessments, followed by a mentor-specific task, an interview, or both.

Applicants will rank their preferred research areas and mentors. Mentors will also rank shortlisted candidates, with both sets of preferences used to determine project matches.

The fellowship is full-time and cannot be completed remotely. SASH says small external commitments may be accepted if applicants disclose them during the selection process.

SASH wrote on LinkedIn: "The fellowship is built for researchers who want their work to inform real-world AI safety practice."

Two online information sessions will take place before the application deadline. The English-language session is scheduled for July 1 at 8 pm Singapore time, while the Mandarin session will take place on July 2 at 5 pm Singapore time.

Mentors span technical safety and AI governance

Fellows will meet experienced researchers each week and receive additional support from dedicated Research Managers.

Confirmed mentors include Soren Mindermann, Scientific Lead for the 2025 International AI Safety Report; Tan Zhi Xuan and Min-Yen Kan at the National University of Singapore; Dong Yinpeng and Xu Mengdi at Tsinghua University; and Pan Xudong at Fudan University.

The mentor group also includes James Chua at Anthropic, Jinge Wang at Concordia AI, Lukas Struppek at FAR.AI, Noam Kolt at Hebrew University, and Konstantinos Krampis at Hunter College, City University of New York.

Mohan Kankanhalli, Director of the National University of Singapore AI Institute and Provost's Chair Professor of Computer Science, will participate with co-mentors Harry Cheng and Wong Yong Kang.

Ryan Kidd, Co-Executive Director of the MATS Program, is advising the fellowship. SASH says additional mentors will be announced before the cohort begins.

Projects may be completed individually or in teams, depending on the mentor and research question. The expected output will usually be a research paper, although other formats may be accepted where they are more appropriate to the project.

Research will cover safety, agents, and loss of control

The fellowship identifies technical AI safety, agent governance, and loss of control as its principal research areas.

Projects will examine how advanced AI systems can be made safer, more secure, and easier to govern. The program will also focus on converting technical research into policy frameworks and recommendations that can be used by institutions.

A second part of the program will connect researchers working across Eastern and Western institutions. SASH says the Singapore location is intended to support collaboration between universities, AI organizations, and policy communities in different regions.

Fellows will work from SASH's office in Singapore's Chinatown district. The program will also include talks, workshops, and events involving mentors, researchers, and people working in AI safety and governance.

Applications remain open until July 10. The selected cohort will begin work in Singapore on September 21 and complete the fellowship on December 4, 2026.