Executive summary: This post reflects on the 2023 iteration of the PIBBSS Summer Fellowship, a 3-month program pairing PhD/postdoc fellows with mentors to collaborate on AI safety research, summarizing key aspects of the program and sharing learnings.
Key points:
The fellowship aims to bring in external expertise to diversify perspectives and methods in AI safety research.
In 2023, there were 18 fellows paired with 11 mentors, plus a 6-week in-person residency and new project report structure.
Both fellows and mentors found value in the collaborations and connections formed through the program.
Research output increased compared to 2022 across a range of AI safety topics according to organizer assessment.
Key successes were transitioning academics to AI risk, high praise for some fellows’ potential impact, and research contributions in interpretability, updateless decision theory, and other areas.
Challenges remain around mentor bandwidth and further improving research output.
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: This post reflects on the 2023 iteration of the PIBBSS Summer Fellowship, a 3-month program pairing PhD/postdoc fellows with mentors to collaborate on AI safety research, summarizing key aspects of the program and sharing learnings.
Key points:
The fellowship aims to bring in external expertise to diversify perspectives and methods in AI safety research.
In 2023, there were 18 fellows paired with 11 mentors, plus a 6-week in-person residency and new project report structure.
Both fellows and mentors found value in the collaborations and connections formed through the program.
Research output increased compared to 2022 across a range of AI safety topics according to organizer assessment.
Key successes were transitioning academics to AI risk, high praise for some fellows’ potential impact, and research contributions in interpretability, updateless decision theory, and other areas.
Challenges remain around mentor bandwidth and further improving research output.
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.