Who's In Charge?
The Governance Crisis Nobody Actually Planned For
The people who understand AI are building it. The people who can regulate it don’t understand it. We are running significantly short of time.
Somewhere in Brussels, a committee is drafting regulations for a technology that will have changed fundamentally by the time those regulations take effect. Somewhere in Washington, a senator is preparing AI questions that will demonstrate they don’t understand how the internet works, let alone what comes after it. Somewhere in San Francisco, a team of engineers is shipping a capability update that renders three months of policy analysis obsolete before it is published. This is AI governance in 2026. It is not a stable situation.
The World’s First Major AI Law: What It Does and Doesn’t
The EU AI Act is the most comprehensive binding legal framework for AI anywhere in the world. It classifies AI by risk level and prohibits the most dangerous applications. This is a genuine achievement. Legislative processes move in years. AI capabilities move in months. By the time a regulation is implemented, it may be addressing yesterday’s system while today’s version has moved beyond its reach.
Legislation moves in years. AI capabilities move in months. By the time a regulation is implemented, it may be addressing yesterday’s system while today’s version has moved well beyond its reach.Neal Lloyd · Inside The Machine, Day 5
Should AI Be Open Source? The Debate That Will Shape Everything
The case for open source: models can be audited, smaller organisations can build on them, geopolitical dependence on a handful of US tech companies is reduced. The case against: a sufficiently capable model, once released, cannot be recalled. If it can assist in designing biological weapons or attacking critical infrastructure, that release is irreversible. Both arguments are serious. Neither is obviously wrong.
The people with the deepest understanding of AI risks are employed by the companies building AI. The people with the power to regulate it are in governments that lack technical expertise. The people most affected have the least influence. This is the governance challenge of the moment.
The Infrastructure Being Built Right Now
Facial recognition AI can identify a specific individual from camera footage in a crowd, in real time, and cross-reference against databases of criminal records, travel history, and financial transactions. This capability exists. It is deployed in 109 countries. Legal frameworks governing its use range from comprehensive and restrictive in a few jurisdictions to essentially nonexistent in most. Good governance requires technical expertise, mandatory transparency, and international coordination. Currently, we have none of these at sufficient scale.
Inside The Machine, Day 05 · May 2026
Neal Lloyd writes about technology, human adaptation, and the uncomfortable questions nobody wants to answer at dinner. Inside The Machine is his ongoing daily series on AI.
- Day 01What Is This Thing?Published — add real URL
- Day 02Survive the MachinePublished — add real URL
- Day 03The Great DebatePublished — add real URL
- Day 04Who Gets Hurt?Published — add real URL
- Day 05Who's In Charge?Published — add real URL
- Day 06The Industries That WinPublished — add real URL
- Day 07The Human EdgePublished — add real URL
- Day 08The Creativity QuestionPublished — add real URL
- Day 09Does AI Feel Anything?Published — add real URL
- Day 10The Data ProblemPublished — add real URL
- Day 11The Trust QuestionPublished — add real URL
