People
Updated 05/18/26Founder
AI safety coordination & strategy collaborator
Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- $35,000
- Funding Raised to Date
- -
Org Details
Updated 05/18/26Horizon Events is the events-focused brand of Horizon Omega (HΩ), a Canadian not-for-profit corporation (registration #1584536-0) based in Montreal, Quebec. The organization was established by Orpheus Lummis, who first organized the AI Safety Unconference at NeurIPS 2018 in collaboration with Effective Altruism Quebec and MILA. The stated mission is to advance AI safety through collaboration, research, and education, with a particular emphasis on reducing catastrophic and existential risks from AI systems. The organization's flagship initiative is the AI Safety Unconference (AISU), a series of participant-driven unconferences held as side events to major ML conferences such as NeurIPS. The format encourages moderated discussions, lightning talks, one-on-one sessions, and breakout groups, prioritizing organic connection and collaboration over formal presentations. Editions were held in 2018, 2019, and 2022 at NeurIPS, and a virtual edition (VAISU) took place in May 2024 with over 400 registered participants, 40 sessions, and 23 recorded talks published on YouTube. A hybrid edition is planned for 2026. The Guaranteed Safe AI Seminars, launched in 2024, are monthly technical seminars featuring leading researchers working on quantitative and high-assurance AI safety approaches, drawing on frameworks outlined in the academic paper 'Towards Guaranteed Safe AI.' The series has over 275 subscribers and has hosted speakers including Steve Omohundro and Rafael Kaufmann. AI Safety in Montreal is a local field-building hub serving over 1,600 members across the Montreal AI safety, ethics, and governance community. The Canadian AI Incident Monitor, launched in 2026, is a bilingual (English/French) structured database tracking AI failures, near-misses, and hazards, with an open API and governance tracking functionality. Horizon Events/Horizon Omega is not a registered charity and does not issue tax receipts. Historically, most work has been self-funded by the founder, with some grant support from the Long-Term Future Fund (LTFF) for the Guaranteed Safe AI Seminars and VAISU, and a grant offer from the Survival and Flourishing Fund for a 2025 workshop. The organization is very small, with Orpheus Lummis as the primary full-time contributor and a network of volunteers and collaborators.
Theory of Change
Updated 05/18/26Horizon Events operates on the premise that well-designed, high-quality events tailored to the AI safety research community are a high-leverage intervention for accelerating safety R&D. By facilitating information flows, building trust and relationships, and enabling collaboration among researchers, the organization aims to increase the quantity and quality of AI safety research output. Events lower coordination costs for researchers, surface new ideas through cross-pollination, and help field-builders and newcomers identify where to contribute. The cumulative effect of stronger research networks and faster knowledge-sharing is intended to increase the probability of solving critical AI safety problems before transformative AI systems are deployed.
Grants Received
Updated 05/18/26Projects
Updated 05/18/26Weekly newsletter that aggregates newly announced AI safety events and training programs worldwide, online and in person.
Local field-building hub and newsletter for the Montréal AI safety, ethics, and governance community, serving 1,600+ members with meetups, coworking, workshops, and collaborations.
Participant-driven unconference series where AI safety researchers propose and run sessions—talks, discussions, and 1:1s—alongside major ML conferences and online to advance AI safety research.
Discussion
No comments yet. Be the first to share your thoughts.