Pivotal fellowship for AI safety research with mentors from GDM, Apollo, Redwood (www.pivotal-research.org)

Pivotal has opened applications for its 9-week Research Fellowship in AI safety and governance (Feb 9–Apr 10, 2026; apply by Nov 30, 2025), offering hands-on mentorship from senior researchers (mentors include contributors from GDM, Apollo, and Redwood), in-person placement at the London Initiative for Safe AI (LISA), and financial support (£6,000–£8,000 stipend, £2,000 accommodation, £2,500 compute, travel/meals). Fellows work full‑time on high‑impact projects with weekly mentor meetings, produce papers or alternative outputs (typical deliverables 10–20 pages), and can apply for extensions up to six months; recent cohorts saw 70% of extension applicants funded for an average of four additional months. Significance: the fellowship is a concentrated pipeline for accelerating AI‑safety talent, bridging early‑career researchers into influential roles across industry, academia, and policy (past alumni moved to DeepMind, GovAI, PhD programs at Oxford/Stanford, founded evaluation orgs, etc.). Practically, the program’s dedicated compute and co‑working resources plus senior mentorship lower barriers for empirically and computationally intensive safety work—enabling reproducible evaluations, red-team studies, governance analyses, and prototype mitigations. For researchers and institutions focused on reducing global catastrophic risk from AI, Pivotal offers a scalable model to seed rigorous projects and long‑term collaborations that can shape safety standards and research agendas.