Lila wants practical chemistry judgments to shape AI safety boundaries
Archived report. Findings and vacancy status have not been reverified.
Lila already describes evaluations for scientific agents and laboratory risks. Its chemistry-safety role asks a chemist to assess practical constraints, expertise and resources, then use those judgments to shape policy boundaries, evaluation priorities and mitigation requirements. As evidence changes, the chemist must distinguish established capabilities from plausible developments and weakly supported hypotheses. Practical chemistry judgment would inform safety priorities; the vacancy does not establish who makes the final decision or whether safeguards work.
The constraint
Scientific capability scores alone do not show which risks become practically achievable when models gain tools and lab access.
Another explanation
The role may supply advice within the existing safety programme rather than determine final controls. No deployed decision, novel safety method or safeguard effect is established.
Evidence
- Lila Sciences · Senior / Principal Chemist, AI Safety
Translate scientific risk judgments into dual-use policy boundaries, evaluation priorities and mitigation requirements that preserve beneficial chemistry.
Context
- Lila scientific AI safety research at MATS
The current first-person research programme already connects scientific-agent and laboratory risks with evaluation coverage and mitigations. Broad integrated safety is already public. The company-specific increment is the chemist responsibility linking practical feasibility judgments to policy boundaries, evaluation priorities and mitigation requirements. Winter 2027 labels the programme, not the page publication date.
What would change this reading?
A harmless, sanitized decision example showing whether a revised feasibility judgment changes or explicitly leaves unchanged an evaluation priority or policy boundary, with the final decision owner identified. Another generic safety announcement would remain inconclusive.
Candidate implications at review
The remit asks chemists to connect feasibility and uncertainty to decisions made by policy, evaluation and mitigation teams, rather than only catalogue possible risks.