Community AI safety day event organized by AI Safety Poland in Warsaw. BlueDot Impact community event bringing together AI safety researchers and practitioners for networking and discussion.
Open deadlines
Conferences, workshops, fellowships, and training programmes with currently-open submission or sign-up deadlines: calls for papers, fellowship applications, registration, early-bird pricing. Sorted by deadline, soonest first.
Regional EA Global (EAGx) conference. Heavy AI safety attendance and programming. In scope for social and community event reasons. Application deadline October 25, 2026.
Structured AI research fellowship designed to help students and professionals publish at premier conferences like NeurIPS, ICML, ICLR, and EMNLP. Two tracks: Fundamental AI Research (AI capabilities, alignment, evaluation, and interpretability) and Applied AI Research (open problems across disciplines). Recent cohorts achieved 68-73% conference acceptance rates.
Annual multi-track interdisciplinary meeting founded in 1987. Includes invited talks, demonstrations, symposia, oral and poster presentations, workshops, tutorials, competitions, and evaluations and datasets track. Major ML conference with AI safety workshops. Three locations: Sydney (main, Dec 6-12), Atlanta (satellite, Dec 9-13), Paris (satellite, Dec 9-13).
Grant fund awarding at least $200,000 in grants and prizes for corrigibility research in 2026. Typical awards range from $5,000-$35,000. Two rounds: Round 1 deadline August 23, Round 2 deadline October 31. Prize awards of $40,000 by end of September and at least $60,000 mid-December (no application required for prizes). Research spanning pure theory (formal models, decision-theoretic analysis) to empirical work (training experiments, evaluations) qualifies. Work must be legible to frontier AI labs and avoid accelerating capabilities development.
Three-week intensive bootcamp for ML upskilling in AI safety focusing on interpretability and RL, based on ARENA curriculum. Covers ML fundamentals, transformer architecture, mechanistic interpretability, LLM evaluations, and alignment science with capstone research project. Fully provided housing, meals (weekday lunches/dinners), 24/7 office access, dedicated TAs, and travel reimbursement up to $750+. Run by Cambridge Boston Alignment Initiative.
4-week intensive program in applied mathematics for AI alignment by Iliad. Based in London at LISA. Provides $5,000 support. Selection based on estimated mathematical strength.
15-week full-time AI safety research fellowship based in London. Expert mentorship through weekly one-on-one sessions with established researchers, dedicated research management support, workspace at LISA with meals on weekdays, financial assistance for travel/housing. Stipend £15,000-£18,000 (£18,000 for senior fellows with PhDs or industry experience). Open to applicants 18+, no specific background required. Remote participation rare but possible.
Foresight Institute Vision Weekend event with frontier science and technology tracks including AI safety. 40-year-old organization focused on transformative technology. Registration open.
Three-month online AI safety research program where participants form teams to work on preselected projects led by Project Leads. Teams meet weekly and plan their own work hours (minimum 10h/week for Project Leads). Final presentations at online conference April 23-26. Project Leads guide research on topics ranging from alignment to stopping harmful AI developments to deconfusion research.
Foresight Institute's flagship event gathering leading scientists, entrepreneurs, funders, and policymakers to explore the frontiers of science and technology, and plan for flourishing futures. Includes AI safety track content. Registration open.
Regional EA Global (EAGx) conference for Australasia. Heavy AI safety attendance and programming. In scope for social and community event reasons. Applications open.
3-month online AI safety research program where participants form teams to work on preselected projects. Project Leads propose and guide research projects across various AI safety topics including alignment, deconfusion research, and projects aimed at stopping or pausing AI development. Teams meet weekly with participants committing 10+ hours per week. Final presentations held at online conference.
Five-month structured training programme on AI evaluation for capabilities and safety. 40 fully-funded scholarships for participants to work on real AI evaluation problems with faculty mentors from Cambridge, Stanford, Princeton, Oxford, MIT, and organizations like Google DeepMind, Microsoft Research, and UK AI Safety Institute. Includes 12 modules, two hands-on courses, and team-based capstone project.
10th annual Cognitive Computational Neuroscience conference bringing together researchers in cognitive science, neuroscience, and AI to explore computations underlying complex behavior. Features keynote speakers, oral presentations, posters, and community-organized sessions including Generative Adversarial Collaborations. Relevant for ACO practitioners interested in predictive-coding, metacognition, and signal-detection-theory measurement work applicable to LLMs.