Abdul El-Sayed releases companion plan to his public-control AI platform

Michigan Democratic Senate candidate Abdul El-Sayed released a companion AI proposal in October 2026, adding AI Under Control to his earlier AI Under Democracy platform. Guardian opinion writer Jerel Ezell described the development in an October 29 column examining El-Sayed’s push to give the public a controlling voice inside influential AI companies—not just rules governing their conduct.

The distinction matters to understanding the agenda: restrictions on AI development would sit alongside changes to who governs the companies developing it. These are campaign proposals, not enacted policy. Ezell’s column does not detail how the October companion changes the earlier platform, so its provisions should not be read as an entirely new package announced this month.

A public stake versus a public vote

El-Sayed’s campaign announced AI Under Democracy on June 29, describing 22 proposals covering ownership, governance, safety and economic protections. Its campaign announcement backed Senator Bernie Sanders’ American AI Sovereign Wealth Fund Act, which would transfer 50% equity from the largest AI companies into public hands. El-Sayed’s stated addition was public control: a seat at the table where company decisions are made.

According to Ezell, the governance proposal calls for 50% plus one of board seats at influential AI companies to be publicly appointed or democratically elected. That is a simple majority, rather than a minority advisory presence. Ezell argues this could give Americans power to block decisions contrary to the public interest, including decisions involving data centers or biological weapons. That is his case for the proposal, not a demonstrated outcome.

Removing corporate influence, adding public safeguards

The agenda also seeks to reduce established technology companies’ influence over AI developers. The June platform lists mandatory divestiture from legacy technology companies. Ezell describes a requirement that tech giants divest from or abandon investments in major AI companies within a defined transition period. He presents those holdings as a conflict between revenue goals and public safety, rather than treating investment alone as evidence of wrongdoing.

The campaign’s June announcement couples those ownership changes with rules for how AI systems are tested and monitored. Its listed measures include mandatory interpretability standards—requirements aimed at making systems more understandable—independent safety testing, biosecurity requirements and mandatory incident reporting. The same platform proposes an annual AI dividend, education investment, and expanded unemployment benefits and wage insurance. Its scope therefore extends beyond preventing dangerous model behavior to sharing benefits and cushioning economic disruption.

Construction rules and limits on releases

Data-center conditions predate both AI platforms. The campaign says El-Sayed released construction terms in January, seeking local job guarantees, energy rules preventing residential rate increases, and environmental protections. Ezell describes his wider approach as slowing development while addressing noise, pollution and public-health effects. Those are conditions on AI’s physical footprint, distinct from the proposed changes inside company boardrooms.

Ezell also says El-Sayed has called for pausing the most powerful AI releases, banning artificial superintelligence and establishing an AI kill switch. The column supports the broader public-control agenda, but acknowledges a political divide: voters could favor tougher regulation or moratoriums without accepting public control of private companies. Its central argument is that regulation and ownership reform are not interchangeable routes to accountability.