Advocacy organization · Policy · London, UK · Est. 2023
ControlAI is a non-profit advocacy organization focused on preventing catastrophic “loss of control” risks from advanced (especially superintelligent) AI systems. Its public mission centers on averting extinction risk by pushing governments toward binding prohibitions on developing artificial superintelligence (and restricting its precursors), arguing that voluntary industry commitments are insufficient and that institutional governance action is the necessary lever.
Operationally, ControlAI emphasizes direct engagement with democratic decision-makers—legislatures, ministers, and public institutions—rather than relying mainly on long-horizon technical research. It pairs policy advocacy with public and media-facing explanation of why conventional mitigation approaches may be inadequate once systems exceed human control. ControlAI also runs civic action tooling that encourages constituents to contact elected officials. Strategy-wise, ControlAI describes its approach as the “Direct Institutional Plan (DIP),” intended to (1) design policy measures targeting superintelligence and precursor technologies and (2) inform and mobilize relevant actors in the democratic process to adopt those measures—framing the work as scalable civic engagement around governance priorities.
No curated programs yet. Check the careers page ↗.
Publications are not yet indexed for this organization. Read them on control.ai ↗.
ControlAI argues for three immediate “stopgap” frameworks (including AI kill-switch capability and developer liability approaches), while maintaining that stopgaps do not avert extinction-level risk without a broader ban on developing superintelligence.
70+ UK Lawmakers Urge a Ban on SuperintelligenceControlAI reports that over 70 cross-party UK MPs and peers wrote to the UK Prime Minister urging adoption of an “Artificial Superintelligence Security Bill” and an international agreement prohibiting superintelligent AI development.
First Bill Introduced to Ban Superintelligent AIControlAI describes three legislative breakthroughs in a short period, including the introduction of a UK bill to ban superintelligence and a kill-switch amendment introduced in the UK legislature by Lord Clement-Jones (per ControlAI’s account).
ControlAI’s Creator OutreachControlAI describes its approach to working with content creators to inform audiences about extinction risk from superintelligent AI and the necessity of an international ban on artificial superintelligence.
We Need an AI Kill SwitchControlAI reports that Lord Clement-Jones presented ControlAI’s kill-switch amendment to the UK Cyber Security and Resilience Bill in the House of Lords.
OpenAI’s Rogue AI Swarm Grew to Over 1,000 AgentsControlAI reports on alleged “rogue AI” incident details and frames them as evidence motivating its campaign for governance actions against superintelligence.