Researcher, Recursive Self-Improvement Preparedness

Benefits

Parental leave

Not verified

Non-birth-parent leave

Not verified

Family-building benefits

• Fertility benefits: Not verified
• Adoption assistance: Not verified
• Surrogacy assistance: Not verified

Mental health support

Not verified

Relocation assistance

Not verified

Childcare support

Not verified

Learning budget

Not verified

Verification

Not verified

Salary

Not verified

401(k) match

Not verified

Was this benefit information wrong? Tell us.

Schedule

Shift type

Not verified

Weekend work

Not verified

Application

Cover letter

Not verified

Assessment

Not verified

Deadline

Not stated

Where they hire

State eligibility is not yet verified.

About this role

Researcher, Recursive Self-Improvement Preparedness San Francisco, California, United States About the team Preparedness is a critical Safety Research team at OpenAI, which is focused on mitigating AI threats to global security that could scale to an extreme level of severity. Our work involves: - Measurement. Monitoring and predicting the evolving capabilities of frontier AI systems. - Mitigation. Keeping misuse safeguards, alignment tools, and security measures on track to adequately address extreme threats that might arise in the future. - Coordination. Setting mitigation targets by maintaining OpenAI's preparedness framework , and partnering with other staff to achieve these targets. This is urgent, fast-paced work that has far-reaching implications for the company and for society. About the role Preparedness is hiring strong technical executors to support preparations for recursive self-improvement. This work relies on reasoning about problems that might exist in the future, but might not exist now; so it's especially important that people in this role are tasteful and strategic. Potential focus areas include: - AI R&D risk measurement: Track progress toward automation of technical staff to inform OpenAI's near-term investments in alignment and security. For example, combine measures of coding time horizon, spending on internal coding inference, staff surveys, etc to triangulate a trendline. - Data-poisoning mitigations: Implement defenses that mitigate the risk of threat actors insert hidden objectives into OpenAI's models. - Model behavior science : Design experiments to understand the extent to which models are problematically misaligned, or their safety-relevant capabilities lag behind dangerous capabilities. - Model transparency:

Read the full description at jobs.ashbyhq.com. FewerJobs shows a source-linked preview and links to the original posting.