Project summary

Horizon Events is a non-profit dedicated to advancing AI safety research and development through targeted events and community initiatives. Our mission focuses on reducing catastrophic and existential risks related to advanced AI systems.

What are this project's goals? How will you achieve them?

Technical events

• Guaranteed Safe AI Seminars

  • 12 seminars from field experts with community Q&A (in 2024 we had Yoshua Bengio, Steve Omohundro, Rafael Kaufmann, Miyazono, etc).

• AI Safety Unconference 2025 (successor to VAISU)

  • 3 days, online; with custom event app, enabling creating and reviewing sessions, poster sessions, schedule voting, matchmaking, chatting, awards. Collaboratively created with community. Focus on x-risk/catastrophic risk, but open to all AI safety works

  • Target: >500 relevant registered participants

• Montréal AI Safety R&D Events: 2h events with technical talk followed by mingling

  • This will complement the AI Safety reading group at Mila, in which Yoshua Bengio participates

Knowledge distribution, community support

• AI Safety Events & Training newsletter (>950 subscribers): Weekly newsletter on upcoming events and training opportunities, sometimes with notes from past events. As of November, we have >950 subscriber.

• Offer events-related help to partners or projects in AI risk reduction, including those with low budgets

• Collaboration with partners (AIGS, Apart Research, Mila/CAISI, etc.)

How will this funding be used?

The funding is to employ and provide tooling to the core team.

On the lower end, it will enable the founder to work part-time on the project, and we may have to reduce our goals. At a higher end, all above projects will be covered, and more team members will be enabled to work on the project.

Who is on your team? What's your track record on similar projects?

Team planned to be active in 2025: founder, Orpheus Lummis. With extra funding, we will enable work by existing but inactive team members or we will grow the team.

Track record

Events successfully executed:

• AI Safety Unconference at NeurIPS (2018, 2019, 2022): Event series that gathered many important AI safety folks

• Virtual AI Safety Unconference 2024 (VAISU), with 400 registered participants, see report

• Guaranteed Safe AI Seminars featuring prominent speakers such as Yoshua Bengio, Steve Omohundro, ...

Community building:

• AI Safety Events and Training newsletter growth to >950 subscribers, ~27K post views

• AI safety reading group at Mila (2018) with David Krueger, ...

Collaborations:

• Events consulting for Atlas Computing

What are the most likely causes and outcomes if this project fails?

The project may fail because of:

• Lack of funding

• Single point of failure with small team, so eg burnout risk or accident risk

Potential outcomes if failed:

• Disruption to established event series

• Reconfiguration or loss of knowledge-sharing platform (newsletter)

• Reduced AI safety event capacity in Montréal

• Gap in technical AI safety community building

What other funding are you or your project getting?

No funding yet for 2025.

In 2024, we were able to mostly self-fund.

Notes

We are happy to discuss. Reply on Manifund, or reach out via email at team@horizonevents.info.

Horizon Events is incorporated as HΩ, a tax-exempt non-profit in Canada (# 15845360). We otherwise support donations via credit card, wire, etc.