Build AI Safety Skills Through Mentorship
Mentee applications are now closed as of January 8th
The next SPAR round will happen later in 2025
We believe that AI might reshape the world in ways that humanity is unprepared for. Therefore, we think ensuring AI's safe development and deployment is a critical priority.
SPAR offers students and early-career professionals the chance to engage in AI safety research through mentorship opportunities in both technical and governance-focused areas. Mentors can obtain additional capacity for their projects, or if more junior, try their fit for leading research.
Our Vision
Information about the program
-
SPAR is a virtual, part-time, volunteer research program designed to give early career individuals and professionals an opportunity to work on a AI safety project for 3 months. Time commitments typically range from 5 to 20 hours per week for mentees and 2-10 hours per week for mentors.
We think its important for mentee and mentors to have legible outputs from the project so we have mandatory midterm and a final report, as well as an optional poster session.
The Spring 2025 round will run from February 10 to May 17. -
What we are looking for in a mentee:Participants who have a technical background (e.g., ML, CS, cybersecurity, math, physics, neuroscience, etc.) or knowledge relevant to policy and governance (e.g., law, international relations, public policy, political science, etc.). The program is open to undergraduate, graduate students and professionals.
Some projects may require additional background knowledge, like knowledge of specific research areas, techniques, or tools (e.g. ML frameworks).
Depending on the project, no prior research experience may be necessary, as long as you meet the specific technical requirements.
What we are looking for in a mentor:We are excited about mentors who are graduate students, faculty members, industry professionals, full-time researchers, and independent researchers with a strong background in AI safety research. We generally don’t accept undergraduate applicants except in exceptional circumstances.
SPAR mentors offer their expertise by proposing, overseeing, and aiding projects on AI safety research. Mentors not only advance their projects but also learn how to mentor others, which we believe is an extremely important and neglected skill in the AI safety community. Through SPAR, mentors gain access to qualified applicants to advance their research proposals, along with operational support and compute funding to help drive their projects forward. Graduate students, faculty members, and professionals with interest are encouraged to fill out a project proposal form, sharing the details of a project they would like to lead.
Being a Mentor
SPAR mentees, from early-career individuals to professionals, take part in AI safety projects over a 3-month period. During this time, they gain hands-on research experience and guidance from mentors, developing skills in safety work while connecting with potential long-term collaborators. Mentees commit 5-20 hours a week to their projects, making the program accessible for those balancing school or work. They also take part in regular check-ins with teams and supervisors, gaining exposure to a collaborative research environment.