Pivotal Research is inviting applications for its AI Safety Research Fellowship 2027, a 15-week, full-time research programme designed to help talented individuals contribute to technical AI safety, AI governance, and related research areas. The fellowship offers expert mentorship, dedicated research management, workspace in London, and financial support to help participants advance projects addressing challenges associated with advanced artificial intelligence.
The programme will run from 18 January to 30 April 2027 at the London Initiative for Safe AI (LISA) in London, United Kingdom. Selected fellows will receive a stipend of £15,000–£18,000, alongside support for travel to London, accommodation for eligible fellows based outside London, and research computing resources.
Applications close on 1 November 2026, making this an opportunity for prospective applicants to prepare their research interests and mentor preferences before the deadline.
Overview of the Pivotal AI Safety Research Fellowship 2027
- Fellowship: AI Safety Research Fellowship 2027
- Organiser: Pivotal Research
- Duration: 15 weeks
- Programme dates: 18 January–30 April 2027
- Location: London Initiative for Safe AI (LISA), London, United Kingdom
- Stipend: £15,000–£18,000
- Accommodation support: £1,800 per month for fellows based outside London
- Additional support: Travel to London and research computing resources
- Work arrangement: Full-time, in-person programme
- Application deadline: 1 November 2026
What Is the Pivotal AI Safety Research Fellowship?
The fellowship aims to expand the number of researchers working on making advanced AI systems safer and ensuring that their development benefits humanity. It provides a structured environment in which participants can develop research ideas, receive feedback from experienced researchers, and make meaningful progress on projects related to AI safety.
Participants can explore technical AI safety, AI governance, technical AI governance, and other relevant research areas. Previous fellowship projects have included evaluating AI agents’ compliance with sanctions law, examining potentially deceptive behaviour in AI models, developing disclosure frameworks for AI laboratory insiders, and investigating privacy-preserving methods for monitoring AI computing infrastructure.
Applicants can indicate interest in working with multiple mentors whose expertise aligns with their proposed research direction.
Fellowship Benefits and Financial Support
1. Stipend of £15,000–£18,000
Selected fellows receive a stipend within the advertised range to support participation in the 15-week programme. The exact amount applicable to an individual fellow should be confirmed with the organiser.
2. Accommodation support
Fellows based outside London are eligible for advertised housing support of £1,800 per month. Applicants should confirm the detailed conditions and how this support is administered.
3. Travel assistance
The fellowship includes support for travel to London. Applicants should consult the official programme information for the applicable arrangements and eligibility conditions.
4. Expert research mentorship
Participants receive weekly one-to-one meetings with established researchers. Mentors provide methodological guidance, critical feedback, and strategic direction to help fellows develop stronger research projects.
5. Dedicated research management
Pivotal’s research managers help participants define achievable project scopes, address research obstacles, maintain progress, and consider their longer-term professional development.
6. Research workspace and resources
Fellows work in person at the London Initiative for Safe AI. Weekday lunch and dinner are included, and participants receive access to the computing resources needed for their research.
7. Workshops and expert discussions
Workshops and question-and-answer sessions with domain experts provide opportunities to develop research skills, explore emerging questions, and establish professional connections.
Who Should Apply?
The supplied programme information does not specify a comprehensive list of formal eligibility requirements, such as mandatory degrees, minimum years of experience, or nationality restrictions. Applicants should therefore consult the official fellowship website and application portal for the complete eligibility criteria.
The programme may be relevant to people interested in technical AI safety, AI governance, AI security, interpretability, AI evaluation, and related fields. Applicants should be prepared to undertake focused, full-time research and explain how their interests align with the work of potential mentors.
International applicants should also confirm UK visa and immigration requirements, as well as any programme-specific conditions governing participation and financial support.
Application Process
Applications are submitted through the official Pivotal Research application portal. The advertised process consists of four stages:
- Written application: Applicants describe their research interests and identify preferred mentors. The estimated time required is one to two hours.
- Video interview: Shortlisted applicants complete a recorded interview lasting less than 10 minutes.
- Research work task: Candidates complete a mentor-specific research exercise, estimated to take one to three hours.
- Personal interview: Applicants participate in a 30–45-minute interview with the team or prospective mentors.
Applicants are encouraged to consider multiple mentors whose research interests genuinely match their own.
How to Strengthen an Application
Applicants should develop a clear explanation of the AI safety or governance problem they want to investigate. They should demonstrate familiarity with relevant research, explain why the problem matters, and identify how their skills and interests could contribute to a practical research project.
Reviewing the profiles of potential mentors and studying previous fellowship projects can help applicants identify suitable research directions. Applications should be specific about mentor preferences and avoid presenting broad interest in artificial intelligence as a substitute for a focused research agenda.
Candidates should also prepare for the work task by practising structured problem-solving, critical analysis, and clear communication of research findings.
Why Consider This Fellowship?
Pivotal Research reports that previous fellows have progressed to organisations and research groups working on AI safety and governance, pursued doctoral studies, founded new initiatives, and published research in academic and policy venues.
These examples demonstrate possible career pathways, although individual outcomes are not guaranteed. For applicants seeking a structured transition into AI safety research, the combination of mentorship, dedicated research management, and sustained project development may provide valuable professional experience.
Application Deadline and Official Application Link
The deadline for the Pivotal AI Safety Research Fellowship 2027 is 1 November 2026. The programme runs from 18 January to 30 April 2027 in London, United Kingdom.
