Meet the team
Gabriel Alfour
CEO and Founder
Gabriel Alfour is CASIR’s Founder and CEO. He is also an advisor of ControlAI, being the originator of its Direct Institutional Plan, and a co-founder of AI safety start-up Conjecture and Marigold, a cryptographic distributed systems R&D company. He leveraged his experience in fundraising for multiple companies to advise ControlAI on its successful fundraising endeavours. He founded CASIR in 2026 to build a well-funded ecosystem of organizations working openly and directly on preventing extinction risk from superintelligent AI.
Eva Behrens
Head of Policy
Eva Behrens is CASIR’s Head of Policy. She is also Special Projects Lead at ControlAI, and was previously Head of Operations and Policy at Conjecture and an AI policy researcher at the Centre for Future Generations, where she authored a policy paper on internationalising AI governance. She is an applied researcher and policy analyst with a focus on aligning economic theory with grand societal challenges, such as climate change and transformative artificial intelligence.
Kemal Evrim
Head of Operations
Kemal Evrim is CASIR’s Head of Operations. He is a serial entrepreneur, co-founder of Digital Village Grand Est and Brave New Media, founder of Templar. He is a seasoned project manager, with more than 18 years of experience in software development management and strategic partnerships. His areas of expertise range from marketing and communication to technical aspects such as handling development teams, AR/VR solutions development, and 3D gaming engines.
Meet our partners
Members
CASIR member organizations are primarily focused on reducing the extinction risk from ASI. CASIR aims to support them through unrestricted funding.
ControlAI is a non-profit that informs policymakers and the public about the extinction risk from artificial superintelligence, and helps them take action to prevent this risk. ControlAI is active in the UK, Canada, US, and Germany, and the team met and briefed over 400 lawmakers in under 2 years across these countries. In the UK, over 160 Members of Parliament support ControlAI’s campaign statement recognizing extinction risk from AI, and over 35 Canadian Members of Parliament support a campaign statement calling for an international “trust but verify” regime prohibiting the development of ASI. In September 2026, ControlAI’s bill to ban superintelligence was introduced in the UK Parliament by Alex Sobel MP, the first bill of this kind to be introduced in any legislature around the world. Over 50,000 people have used ControlAI’s tools to send over 200,000 messages in total to policymakers, asking them to take action to prevent ASI extinction risk.
PauseAI Global is a non-profit created to build the visible public support that policymakers need to take action on preventing ASI extinction risk. PauseAI Global is a federation active in 15+ countries, with semi-autonomous local chapters, united by the shared goal of a binding international pause on frontier AI development. Among other things, the Federation held a conference at the European Parliament, downstream of which Brando Benifei, co-rapporteur of the EU AI Act, proposed that the GPAI Code of Practice be extended to include conditional restrictions on development beyond capability thresholds. In France, PauseAI has achieved over 1 million cumulative media views over 40+ appearances, and held the largest Francophone AI safety conference to date. In the UK, 60 policymakers from 10+ parties signed PauseAI UK’s open letter to Demis Hassabis, and PauseAI has held the largest AI safety protest held anywhere to date, with 300 people in London in February 2026.
The Existential Risk Observatory was founded to spread awareness about existential risks, including extinction, from AI and other technologies. The ERO has hosted public events with panel discussions on AI risks and policy solutions, and its team members have published research on how international agreements on AI can be implemented practically. ERO team members have been featured in Time Magazine and leading Dutch newspapers, such as De Telegraaf, NRC, and Trouw.
PauseAI US is a nationwide grassroots organization working to pause the development of frontier AI, and calling for the US to lead an international treaty to pause AI development. PauseAI US supports local groups across the United States that meet with lawmakers, hold public protests, and engage the media to call for such a pause. As part of this work, PauseAI US provides volunteers with free grassroots lobbying training and workshops. Its PauseCon events bring together volunteers from all over the US in Washington D.C. to meet their elected representatives in person.
Evitable was founded by AI professor David Krueger to inform and organize the public around societal-scale risks and harms of AI. Projects include AI on the Ballot, a nonpartisan, interactive website tracking the stances of U.S. Congressional candidates in the 2026 elections on AI, and numerous media appearances, including in USA Today, the Washington Post, the Economist, and the Guardian.
The Torchbearer Community is a network of volunteers working to prevent extinction risk from ASI by informing the public and providing it with tools to take action. Its members come from all walks of life, including other CASIR partner organizations. Torchbearer community members have run projects that have led to hundreds of lawmakers across multiple countries being contacted, asking them to address extinction risk from ASI, and have built tools that help citizens contact their lawmakers and urge them to take action. Furthermore, the community has supported multiple members in founding and running local volunteer and action groups, including PauseAI Global’s Germany chapter.
Partners
CASIR Partner organizations have extinction risk from ASI as one of their topics of interest. CASIR may fund them for specific projects that follow the Charter.
CeSIA is an independent non-profit organisation and think tank dedicated to preventing major risks associated with AI by informing policymakers, raising public awareness, and conducting policy research. CeSIA partners with the European AI Office and the UK AI Security Institute on evaluating AI models, and initiated an international call for red lines to prevent unacceptable AI risk, signed by 15 Nobel Prize and Turing Award recipients, and over 300 notable figures in total. At the national level in France, CeSIA has advised the Ministry of Defence, CNIL (France’s national data protection agency), and other institutions on AI policy, and has launched 3 accredited courses on AI risk at leading French universities.
The Alliance for Secure AI works to build enforceable safeguards to ensure AI is developed and deployed under meaningful human control, in the public interest, and with accountability when it causes harm. Key goals include educating the public about the risks of advanced AI, and building bipartisan support in the US for enforceable safeguards on AI. The Alliance hosts panels and debates on the impact of AI at conferences across the country, convenes faith leaders on AI, hosted a screening of “The AI Doc”, a documentary about AI risk, for Members of Congress and their staff, and runs a website tracking the loss of American jobs due to AI at jobloss.ai .