Programme Thesis
The OpenPhil Career Transition fellowship provides up to $100K to enable researchers to pivot into high-impact AI safety or biosecurity work by funding a period of exploration and capacity-building. It exists to attract talented individuals from adjacent fields into areas where Open Philanthropy believes additional research effort is most needed.
Selection Criteria
- Applicant must demonstrate a credible plan to transition into AI safety or biosecurity research, with clear pathways to impact.
- Proposed research should align with Open Philanthropy's priorities (e.g., AI alignment, catastrophic risk reduction, biosecurity).
- Applicant should have a strong track record in a relevant technical field (e.g., ML, neuroscience, pharmacology, statistics).
- Evidence of institutional support or collaboration with established researchers/groups in the target area is highly valued.
- The proposal should show high leverage: co-investment, existing networks, or potential to build capacity in underserved regions.
- Personal statement must convey genuine commitment to the cause and a realistic theory of change.
Past Winners / Cohort Profiles
The page describes a single unsuccessful application from a mid-career AI methods researcher embedded in a large organization (CSIRO). No named winners are provided. Archetypes likely include early-to-mid-career researchers with strong quantitative skills and a clear pivot plan into AI safety, often with institutional backing.
Ideal Candidate Fingerprint
A technically strong researcher (e.g., in ML, statistics, neuroscience, or pharmacology) with a concrete, well-scoped plan to transition into AI safety or biosecurity, backed by a reputable institution or collaborator. The applicant should demonstrate both domain expertise and a credible theory of change for how their work will reduce catastrophic risk.
Recommended Framing
Eniola should frame their CCT model and neurocascade simulation engine as foundational tools for understanding and potentially intervening in addictive behaviors—a form of cognitive bias and loss-of-control that parallels AI safety concerns about gradual loss of control. Emphasize how their multi-domain computational skills (ODE modeling, Bayesian calibration, protein ML) can be directly applied to AI alignment problems, and leverage their upcoming M.Sc. at HPI/Potsdam as institutional credibility. Highlight the Africa angle as a unique capacity-building opportunity in a region with minimal AI safety research.
Watch Out
The programme appears to prioritize AI safety and biosecurity; Eniola's current work is in addiction neuroscience and pharmacology, which may be seen as tangential unless explicitly connected to AI safety. The unsuccessful example suggests Open Phil may be skeptical of applicants without a direct AI safety track record or institutional home in that field. Eniola's independent researcher status and lack of a PhD could be a disadvantage against more established candidates.
Research History
2026-07-24 07:28 · medium confidence