ControlAI 2025 Impact Report

ControlAI 2025 Impact Report 5 min read • ControlAI's mission • We often meet lawmakers who had never heard of the risks of superintelligence before our meeting, and walk out with them supporting our campaign. • Our Results Last Year • Lawmaker outreach • Media & content creator outreach • Public awareness campaign and lawmaker engagement tools • Theory of Change • The awareness gap • Building this awareness at scale, among both decision makers and the public, is the necessary first step for any meaningful action on superintelligence. • Building the coalition • These measures can be achieved by a powerful coalition of countries that understands superintelligence as a vital threat to their national security, and treat its development as they would any other severe security threat. • ControlAI exists to make sure that a strong coalition of countries rises up to the challenge of preventing the development of superintelligence. • Our Theory of Change in more detail • Moving forward • In little more than a year, we have proven that directly engaging democratic institutions on the extinction risk from superintelligence works. AI GovernancePoliticsAIWorld Optimization Personal Blog 20 ControlAI 2025 Impact Report by Andrea_Miotti, Alex Amadori 27th Mar 2026 Linkpost for controlai.com 5 min read 0 20 This post highlights a few key excerpts from our full impact report. You can read the full report at https://controlai.com/impact-report-2025. ControlAI is a non-profit organization working to avert the extinction risks posed by superintelligence. We help hundreds of thousands of people understand these risks and meet hundreds of lawmakers to inform them, without mincing words, about what is at stake. In little more than a year, we briefed over 200 parliamentarians, built a coalition of 110+ UK lawmakers recognizing superintelligence as a national security threat, led to two debates in the UK House of Lords, and our work led to a series of hearings on AI risk and superintelligence at the Canadian Parliament.[1]These hearings included testimonies from me (Andrea) and Samuel at ControlAI, Connor Leahy, Malo Bourgon (MIRI), Max Tegmark and Anthony Aguirre (FLI), David Krueger, and more. The report covers results between December 2024 and January 2026. As of posting this in March 2026, we've now briefed 279 lawmakers and 90+ US congressional offices. In only the last 2 months, we've scaled in Canada and Germany from ~50 to 100+ lawmakers briefed, despite us only having one staffer in each country. Moving forward, we plan to significantly expand our work in the US, accelerate our progress from awareness to policy action and establish a presence in all other G7 countries. ControlAI's mission ControlAI’s mission is to avert the extinction risks posed by superintelligence. Nobel Prize winners, top AI experts and even the CEOs of major AI companies have warned that superintelligence poses an extinction risk for humanity. Yet, most decision-makers and most of the public are still in the dark about these risks. To avert these risks, we need to prevent the development of superintelligence. We tackle this problem in a direct and straightforward way: meet all relevant actors in the democratic process, inform them of the risks and the solutions, and ask them to take action on this issue, systematically and repeatedly. We help hundreds of thousands of people understand the extinction risk posed by superintelligence, and to take civic action. We meet hundreds of lawmakers to inform them, without mincing words, about the risks of superintelligence. We help lawmakers speak out publicly and push for concrete measures to prevent superintelligence, nationally and internationally. We often meet lawmakers who had never heard of the risks of superintelligence before our meeting, and walk out with them supporting our campaign. In the UK, we started out with cold outreach to all lawmakers, and one year later we have a coalition of over 110 lawmakers recognizing superintelligence as a national security threat, which has already led to two debates in the UK parliament on superintelligence and extinction risk. We are now expanding our model to other countries, including the US, Canada and Germany. The early results, which we present in this document, show that our playbook can be replicated across borders. We produced these results with a team of fewer than 15 people, operating on a small budget compared to the scale of the problem. Our methods have significant room to scale with more resources. Our Results Last Year Lawmaker outreach ~1 in 2 UK lawmakers we brief go on to support our campaign 110+ UK lawmakers supported our campaign 2 Parliamentary debates on superintelligence and AI extinction risk Media & content creator outreach 18 Media publications on risk from superintelligent AI resulting from our work 14 Videos published in collaboration with content creators totaling 20+ million subscribers Public awareness campaign and lawmaker engagement tools 160,000+ Messages sent to US and UK lawmakers from constituents about superintelligence extinction risk 30,000+ People who contacted their lawmakers through our tools in the US and UK Theory of Change The awareness gap In order to establish strong international coordination to prevent superintelligence, countries will need resources like funding, political will, diplomatic leverage, and sustained attention from capable people. Only deep awareness of superintelligence and its risks will justify such investments in the eyes of these actors. Without genuine understanding and conviction, individuals and countries will not bear the real costs needed to solve the problem, nor remain vigilant as AI development evolves and political or economic circumstances change. Right now, that awareness barely exists. Decision makers and the public across countries are largely unaware of the extinction risk it poses. When we started, virtually no one was bringing the extinction risk from superintelligence directly to lawmakers. Building this awareness at scale, among both decision makers and the public, is the necessary first step for any meaningful action on superintelligence. In order to kickstart the kind of international coordination needed to prevent superintelligence from being built, we need to rally a critical mass of countries that take the risks of superintelligence at least as seriously as they take the threat of nuclear war today, and that treat its development as they would any other severe threat to national security. Building the coalition With sufficient buy-in, the next steps become possible. Informed governments backed by public demand for action can pursue concrete policy measures: national legislation prohibiting the development of superintelligence, and international agreements modeled on existing nonproliferation and WMD-prevention frameworks. These measures can be achieved by a powerful coalition of countries that understands superintelligence as a vital threat to their national security, and treat its development as they would any other severe security threat. Such a coalition has a wide range of enforcement tools available, from formal agreements and inspection regimes to sanctions and multi-lateral monitoring mechanisms, the same tools that have been used to constrain nuclear proliferation and other global security threats. Both superpowers and middle powers are well-positioned to join this effort, as they all face the universal extinction threat from superintelligence being developed. ControlAI exists to make sure that a strong coalition of countries rises up to the challenge of preventing the development of superintelligence. Our Theory of Change in more detail This chart describes our theory of change in more detail, and clarifies how our work fits in it. Moving forward In little more than a year, we have proven that directly engaging democratic institutions on the extinction risk from superintelligence works. The UK is our proof of concept; we are now replicating this model in the US, Canada, and Germany. In the UK, where we already helped move the issue of superintelligence into the halls of politics, 2026 will be the year to translate this momentum into concrete policy change. As we scale, we are confident that more resources will translate directly into more countries where lawmakers understand and act on this threat. We will expand our work in the US, accelerate our progress from awareness to policy action, and establish a presence in all other G7 countries. If you are a donor or partner who wants to help build the coalition that keeps humanity in control, please get in touch at partners@controlai.com. As of March 2026, a member of our coalition has submitted an amendment to a UK cybersecurity bill recognizing superintelligent AI as "systems that can autonomously compromise national security, escape human oversight, and upend international stability". This amendment will be discussed in the UK Parliament. ↩︎ 1. As of March 2026, a member of our coalition has submitted an amendment to a UK cybersecurity bill recognizing superintelligent AI as "systems that can autonomously compromise national security, escape human oversight, and upend international stability". This amendment will be discussed in the UK Parliament. ↩︎ AI Governance11Politics3AI4World Optimization3 Personal Blog New Comment Normal Insert Type here! Use '/' for editor commands. Submit Moderation Log More from Andrea_Miotti 65The Compendium, A full argument about extinction risk from AGI adamShimi, Gabriel Alfour, Connor Leahy, Chris Scammell, Andrea_Miotti 1y 13 50AGI in sight: our look at the game board Andrea_Miotti, Gabriel Alfour 3y 12 30A Narrow Path: a plan to deal with AI extinction risk Andrea_Miotti, davekasten, Tolga 1y 0 View more Curated and popular this week 25A Toy Environment For Exploring Reasoning About Reward jenny, Bronson Schoen 3d 1 20Test your best methods on our hard CoT interp tasks daria, Riya Tyagi, Josh Engels, Neel Nanda 2d 0 0Comments 0 x