Executive summary: The ML Alignment & Theory Scholars program (MATS) successfully ran its fifth iteration in Winter 2023-24, providing mentorship and support to 63 AI safety research scholars, and plans to make several improvements for future programs based on scholar and mentor feedback.
Key points:
Key changes from the previous program included reducing the scholar stipend, transitioning to Research Management, using the full Lighthaven campus, and replacing Alignment 201 with AI Strategy Discussions.
Scholars were highly likely to recommend MATS (9.2/10 average rating, +74 NPS) and rated mentorship highly (8.1/10 average). Mentorship was the most valuable MATS element for 38% of scholars.
Mentors were also likely to recommend MATS (8.2/10 average, +37 NPS). The most common mentoring benefits were helping new researchers, gaining mentorship experience, and advancing AI safety.
According to mentors, 77% of evaluated scholars could achieve a top conference paper, 41% could receive a job offer from an AI safety team, and 16% could found a new AI safety org within the next year.
After MATS, scholars reported facing fewer obstacles to an AI safety career, with the biggest remaining obstacles being publication record and funding.
Key planned changes for future programs include introducing a mentor selection advisory board, shifting research focus, supporting more AI governance mentors, expanding applicant pre-screening, and modifying the strategy discussion format.
This comment was auto-generated by the EA Forum Team. Feel free to point out issues with this summary by replying to the comment, andcontact us if you have feedback.
Executive summary: The ML Alignment & Theory Scholars program (MATS) successfully ran its fifth iteration in Winter 2023-24, providing mentorship and support to 63 AI safety research scholars, and plans to make several improvements for future programs based on scholar and mentor feedback.
Key points:
Key changes from the previous program included reducing the scholar stipend, transitioning to Research Management, using the full Lighthaven campus, and replacing Alignment 201 with AI Strategy Discussions.
Scholars were highly likely to recommend MATS (9.2/10 average rating, +74 NPS) and rated mentorship highly (8.1/10 average). Mentorship was the most valuable MATS element for 38% of scholars.
Mentors were also likely to recommend MATS (8.2/10 average, +37 NPS). The most common mentoring benefits were helping new researchers, gaining mentorship experience, and advancing AI safety.
According to mentors, 77% of evaluated scholars could achieve a top conference paper, 41% could receive a job offer from an AI safety team, and 16% could found a new AI safety org within the next year.
After MATS, scholars reported facing fewer obstacles to an AI safety career, with the biggest remaining obstacles being publication record and funding.
Key planned changes for future programs include introducing a mentor selection advisory board, shifting research focus, supporting more AI governance mentors, expanding applicant pre-screening, and modifying the strategy discussion format.
This comment was auto-generated by the EA Forum Team. Feel free to point out issues with this summary by replying to the comment, and contact us if you have feedback.