Existential Risk Observatory

The Existential Risk Observatory (ERO) is a Dutch foundation (stichting) incorporated in Amsterdam in May 2021, with ANBI tax-exempt status, dedicated to reducing existential risks to humanity by informing the public debate. Its central focus is artificial intelligence; it also addresses man-made pandemics, nuclear war, and climate change as global catastrophic risk categories. The organisation's theory of change distinguishes it from primarily research-oriented x-risk bodies: it treats public communication as a specifically neglected lever — its founding hypothesis was that widespread public awareness of existential risk would generate more talent, funding, research capacity, and policy priority than comparable inputs spent on research or policy work alone. ERO accordingly operates at the public-debate end of the AI safety movement: large events, media engagement, voting guides for elections, and policy submissions to political parties, rather than original safety research.

Founding and structure

Otto Barten, a physicist and sustainable energy engineer, founded ERO in May 2021 in Amsterdam after concluding that existential risks were even more important than climate change — the field he had previously worked in. The EA Forum launch post noted explicit alignment with the Effective Altruism ecosystem and openness to cooperation with the Future of Life Institute, StopAI, and PauseAI. ERO is legally a stichting (non-profit foundation) under Dutch law with ANBI (tax-exempt charitable) status and is headquartered at Havikslaan 8A, Amsterdam. The board consists of Barten as director, Joep Sauren (treasurer, Industry 4.0 specialist), and Marko van der Wal (secretary, editor and translator). Research and policy staff include Hugo Save (AI evaluations), Raymond Koopmanschap (international AI agreements), Soumya Sankar (mathematics postdoc), and Sue Anne Wong (policy researcher). In 2024 the Campaign for AI Safety merged into ERO, bringing its founder Nik Samoylov as senior campaigner — expanding the organisation's public-mobilisation capacity. ERO was self-funded at founding and received a grant from the Long-Term Future Fund and a contribution from an individual donor in January 2024.

Public debate and events

ERO's most visible output has been a series of high-profile public debate events in the Netherlands and beyond. It has hosted debates at Pakhuis de Zwijger in Amsterdam — an established Dutch public-debate venue — on AI existential risk, with keynote speakers including Stuart Russell and Yoshua Bengio. On 9 February 2025, the evening before the Paris AI Action Summit, ERO co-organised an AI Safety Debate at CNAM in Paris featuring Bengio alongside Time technology correspondent Billy Perrigo, The Economist deputy executive editor Kenneth Cukier, Skype co-founder Jaan Tallinn, Dutch Ministry of Foreign Affairs official Emma Verhoeff, and PauseIA founder Maxime Fournes. The event was opened by Art de Blaauw, Government Chief Information Officer of the Netherlands. This event format — convening researchers, journalists, policymakers, and AI safety advocates for structured public debate — is ERO's primary model for engaging general audiences outside the existing x-risk community.

Policy work

ERO has developed a 15-measure AI safety policy platform organised under three pillars: Safety (AI pause implementation, licensing regimes, model evaluations, third-party auditing, hardware tracking), Democracy & Openness (public funding for AI safety research, democratic processes and referendums on AGI development), and Governance (an international AI Agency, liability frameworks, copyright protections). Ahead of the November 2023 Dutch general elections, it produced a stemwijzer (voting guide) analysing the positions of Dutch political parties on AI existential risk — a tool in the Dutch civic tradition of assisting voters in comparing party positions — and submitted policy papers to Dutch parties under the title Veilige & Menselijke AI (Safe and Humane AI). Research topics pursued by the team include the Conditional AI Safety Treaty and coordination mechanisms for a halt on frontier AI development.

Place in the movement

ERO occupies a distinct position in the Dutch and European AI safety landscape. Where PauseAI organises volunteer-led protest chapters and direct action, and ControlAI runs an inside-game parliamentary strategy, ERO focuses on raising the public salience of existential risk through media, debate, and civic-democratic tools — making it the public-debate end of the same Dutch/European AI safety ecosystem. Its absorption of the Campaign for AI Safety in 2024 added mass-mobilisation capacity and reflects a consolidation within the small field of European AI safety advocacy organisations. The academic endorsements from University of Groningen professors in philosophy and computer science add some university credibility to an organisation whose primary posture is civic communication rather than research.