There is and will continue to be mass confusion around AI regulations and regulatory compliance because no one any longer believes that regulators actually “mean it.”
Let me explain.
Over and over, bright, generally young legislators at every level of government have decided that they will regulate AI. This often follows the “a thing must be done, this is a thing, therefore it must be done” logic.
Then, reality intrudes and the law is either pushed back or softened. I think some would like to believe that this is due to regulatory pressure from Elon Musk or whatever — but more often it’s simply legislators figuring out what the bill they passed actually said.
Boy — there are more examples than counter-examples.
New York City Local Law 144. This was the granddaddy of AI regulation. According to a Cornell study, <5% of businesses to whom the law applies actually did so. The key reason? NYC has completely abdicated enforcement of the law. As far as I can tell, there has NEVER been an enforcement action.
Colorado AI Act. Signed grudgingly by Gov. Polis. Implementation pushed back and substantially weakened. This one probably is due to industry pressure, when it became absolutely clear that this would be the end of the tech industry in Colorado.
EU AI Act. Pushed back and substantially ruined. And if you believe the August 2026 date for high-risk systems is going to hold, I have a bridge over the river Rhine to sell you.
The EU AI Act has been a complete comedy of errors, as compliance involves filing paperwork for which there is no template, with an office that doesn’t exist, alongside harmonized standards which have not been identified.
Let me know — what regulations are actually being enforced well?
About the Author
John Rood is the founder of Proceptual, where he helps organizations build practical AI governance systems. He’s an ISO 42001 Lead Auditor and lead instructor for Michigan State University’s AI Governance Certification. His writing has appeared in HR Brew and Tech Target.
