AIxBio projects on biological data, biological AI models, and their interactions with GPAIs, to directly progress the maturity of interventions and reduce strategic/technical uncertainties. Some projects may be scoped to directly inform Sentinel's AIxBio strategy and resource allocation.
Topic areas:
- Securing biological data to reduce risks from biological AI models
- Reducing risks from biological AI models
- Reducing risks arising from the interactions between GPAIs and biological AI models
Objective types:
- Increase policy readiness of a policy intervention (e.g. risk-based governance of bio data)
- Increase readiness of a technical or process intervention (e.g. new bio AI model input/output classifiers)
- Reducing uncertainties that block technical/policy progress (e.g. how bio data links to dual use capabilities, the potential role of synthetic data)
- Increase international diffusion of progress to key jurisdictions
- Achieve buy-in for promising interventions from new and influential stakeholders
- Improve situational awareness (e.g. mapping future data assets, reducing blind spots like proprietary datasets)
I may update later with more specific projects.
AIxBio projects - technical or policy. Priorities include the security of new and existing biological data, the security of large biological AI models, AI creation of biological AI tools, strategy for a world with biological superintelligence.
Light-touch (.5 hours of weekly 1:1s)
London
Strong preference
Indifferent
Toby Webster is a program director at Sentinel Bio, leading work on risks at the intersection of AI and engineering biology. Previously, Webster studied biological AI model security at RAND Europe.
A range of skill mixes is appropriate for the types of work I am interested in, including both technical and policy.
I would prefer a fellow who can carry out independent proposals, consultation, and dissemination. I am not able to give detailed research and project management input.