OpenMined founder reframes AI safety debate around unilateral control

On September 14, Andrew Trask (username iamtrask), founder of OpenMined and former OpenAI researcher, laid out in a tweetstorm his reframing of the AI safety debate: the open-source vs. closed-source fight isn't the real issue—the core problem is "unilateral control over AI."

Confirmed

iamtrask argues that 99.999% (also written as 99.9% in some posts) of the AI safety debate revolves around two questions: "who should unilaterally control superintelligent AI" and "how to ensure AI doesn't unilaterally control itself."
He frames unilateral control in three forms: closed-source AI is unilaterally controlled by one company; open-source AI is unilaterally controlled by anyone holding a copy; and misaligned AI is unilaterally controlled by the AI itself.
He contends that no one in the world can be trusted to unilaterally control superintelligence, and the only answer is "collective control"—which he calls the secret mechanism behind democracy and free markets.

Why it matters

This view recasts the long-standing "open source vs. closed source" divide as a question of power structures, offering a new analytical framework for the debate.
The call for collective control extends OpenMined's consistent decentralized approach to safety, with direct relevance to discussions on superintelligence governance.

2026-09-14 ~ 2026-09-14 · 7 related posts

Primary sources

iamtrask: The Real AI Safety Problem Is Unilateral Control, Not Open vs Closed — iamtrask · 2026-09-14 07:44 (UTC+08:00)
iamtrask: Collectively Controlled AI Is Easier to Regulate With Less Regulatory Capture — iamtrask · 2026-09-14 07:46 (UTC+08:00)
iamtrask: Model Fusion and Ensembles Beat Frontier AI on Accuracy and Cost — iamtrask · 2026-09-14 07:45 (UTC+08:00)

iamtrask: Open vs closed source isn't the issue — unilateral control of superintelligence is — iamtrask · 2026-09-14
[source] iamtrask: The Real AI Safety Problem Is Unilateral Control, Not Open vs Closed — iamtrask · 2026-09-14
iamtrask Maps Three Forms of Unilateral AI Control: Company, Copy Holder, the AI Itself — iamtrask · 2026-09-14
OpenMined's Trask: AI safety debate boils down to who unilaterally controls superintelligence — iamtrask · 2026-09-14
iamtrask: Nobody Can Be Trusted With Unilateral Superintelligence Control — Only Collective Control Works — iamtrask · 2026-09-14
[source] iamtrask: Model Fusion and Ensembles Beat Frontier AI on Accuracy and Cost — iamtrask · 2026-09-14
[source] iamtrask: Collectively Controlled AI Is Easier to Regulate With Less Regulatory Capture — iamtrask · 2026-09-14