AI Safety Accelerator

The platform solving the top 10 existential AI threats.

Better Societies convenes the world's leading AI safety researchers, founders, policymakers, and investors to tackle existential AI threats head-on. 3 programs. 1 mission.

Get started Join the community

Our experts come from, and featured in

Backed by real expertise, not marketing.

Better Societies operates with 10+ confirmed AI safety experts from institutions like Yale, IEEE, Bristol Robotics Lab, UT Austin, and Sorbonne. Every program is shaped by the people defining AI safety standards globally.

Confirmed experts 10+ AI safety researchers, policy advisers, and founders from Yale, IEEE, Bristol Robotics Lab, and more.
Community members 92+ Founders, researchers, policy advisers, and investors building safer AI.
Active programs 3 EU AI Act compliance, AI Safety Accelerator, and the AI Safety Summit.

The top 10 existential threats of AI.

Each card represents one of the 10 existential threats Better Societies works to address. Click to read the full report.

01

Loss of control

AI acting outside intended objectives once deployed at scale.

Misalignment 02

Autonomous weapons

AI-driven weapons making lethal decisions without human review.

Misuse 03

AI-enabled bioweapons

AI lowering the barrier to creating biological threats.

Misuse 04

AI cyber warfare

AI-augmented offensive cyber capabilities targeting critical infrastructure.

Security 05

Mass unemployment

Rapid labor displacement outpacing economic adjustment mechanisms.

Systemic 06

AI surveillance

AI scaling state and corporate surveillance beyond historical limits.

Concentration 07

AI misinformation

Generated content eroding public information integrity and trust.

Misinformation 08

Recursive self-improvement

AI improving its own capabilities faster than human oversight can react.

Misalignment 09

AI-accelerated climate

Compute demand and AI-driven extraction worsening environmental trajectories.

Systemic 10

AI space-based threats

AI-controlled space assets introducing new categories of strategic risk.

Misuse

Source: Slattery et al. The AI Risk Repository, arxiv:2408.12622. Categories from the MIT Domain Taxonomy.

3 ways Better Societies helps.

Whether you're building a startup, advising policy, or ensuring your AI is compliant, there's a program built for you.

Accelerator

AI Safety Accelerator

A fully virtual, global 12-week program that takes you from idea to launched AI safety project. Live curriculum, expert mentorship, funder introductions, and a global cohort of founders, researchers, and policymakers.

  • Connect with funders, co-founders, and advisors from day one
  • Hands-on incubation from idea through first fundraise or revenue
  • Fully remote and open to founders anywhere in the world, non-profit and for-profit alike
  • Live weekly masterclasses and workshops with leading AI safety experts
  • Direct introductions to investors and partners actively funding AI safety
Apply to the Accelerator
Virtual summit

AI Safety Summit

Ongoing virtual summit with the world's leading AI safety experts, including Roman Yampolskiy, Wendell Wallach, Alan Winfield, and 10+ others.

  • Live masterclasses from confirmed AI safety experts
  • Deep-dive discussions on the top existential threats
  • Frontline updates from the people defining AI safety standards globally
  • Free to attend: open to researchers, founders, policymakers, and investors
Explore the Summit
Done-for-you

EU AI Act Compliance

2 August 2026 is binding. Done-for-you documentation and conformity assessment for companies building high-risk AI systems.

  • Annex III high-risk systems face binding obligations from 2 August 2026
  • Done-for-you Annex IV technical file, risk management system, and oversight policy
  • Regulator-acceptable documentation at a fraction of Big Four firm rates
  • 12-month audit defense and satisfaction guarantee included
Get compliant

10+ confirmed AI safety experts.

Each expert shapes Better Societies programs directly. Their research defines the standards AI systems must meet globally.

Roman Yampolskiy

Roman Yampolskiy

Univ. of Louisville

AI safety and controllability research that anchors our loss-of-control threat curriculum.
Wendell Wallach

Wendell Wallach

Yale Center for Bioethics

AI governance and moral machine ethics, frames Better Societies' policy and compliance work.
Alan Winfield

Alan Winfield

Bristol Robotics Lab / IEEE

Robot ethics standards author, brings IEEE-level rigor to our Accelerator's ethics framework.
Raja Chatila

Raja Chatila

Sorbonne / IEEE

Human-robot interaction and AI ethics, shapes how Better Societies defines trustworthy AI.
Risto Miikkulainen

Risto Miikkulainen

UT Austin / Cognizant

Neuroevolution and AI cognition research, advises on recursive self-improvement risks.
Toby Walsh

Toby Walsh

UNSW Sydney

Leading voice on autonomous weapons bans, advises our policy and activist program tracks.
John C. Havens

John C. Havens

IEEE Global Initiative

Author of IEEE's Ethically Aligned Design, co-designed the Summit's governance curriculum.
Alex Polyakov

Alex Polyakov

Adversa AI

Adversarial AI and red-teaming specialist, informs our AI cyber warfare and misuse threat research.
Olaf Witkowski

Olaf Witkowski

Cross Labs

AI emergence and complexity science, advises on systemic risk scenarios for the Accelerator.
Kathleen Richardson

Kathleen Richardson

De Montfort University

AI ethics and social impact research, grounds our program's human dignity and rights framing.

Which one fits you?

Tell us what you're working on and we'll route you to the right program.

Accelerator

You want to build an AI safety project with a global virtual cohort, expert mentorship, live curriculum, and funder access.

Pre-idea or early-stage, non-profit or for-profit, the Accelerator is built for founders and researchers advancing AI safety. Network first, curriculum second.

Apply to the Accelerator
Summit

You want to learn from the world's leading AI safety experts.

The AI Safety Summit brings live masterclasses and deep-dive discussions. Free to attend for researchers, founders, policymakers, and investors.

Explore the Summit
Investors

You invest and want deal flow in AI safety.

Get demo-day access and warm introductions to vetted founders coming through the Accelerator, plus a front-row read on where AI safety is heading.

Become a funding partner
Government and policy

You shape AI policy, standards, or oversight.

Partner with us on AI safety standards, the Summit policy track, and the Top 10 threats research, with the field's leading experts at the table.

Explore a partnership

The AI safety community that works with you, not at you.

Join 92+ founders, researchers, policymakers, and investors building safer AI. Access peers, live sessions, and expert masterclasses from the day you join.

Peers who get it

Connect directly with founders, researchers, policymakers, and investors working on the same problems.

Expert masterclasses

On-demand sessions from confirmed AI safety experts. New content each month from the people defining global standards.

Weekly live call

Join the weekly community call. Ask questions, share progress, and get unfiltered input from the group.

First value in minutes

The moment you join, you get access to our AI Safety Quickstart Guide: a concise map of the top 10 existential AI threats and the concrete actions your team can take this week.

Join the community free

Get started today

Help accelerate AI safety.

Tell us who you are and what you're working on. We'll route you to the right program.

Tell us about you Join the free community