For years, the European Union was the gold standard for AI regulation. The AI Act. The GDPR. The Digital Services Act. Brussels wrote the rulebook that the rest of the world either copied or complained about. It was the one place where Big Tech could not operate without guardrails.

That era is ending. And Amnesty International just published the receipts.

In a detailed analysis released this week, Amnesty lays out how the European Commission's "Digital Omnibus" package, launched under the banner of "simplification" and "cutting red tape," is systematically weakening the AI Act and GDPR in ways that directly benefit the companies those laws were designed to restrain.

The changes are not subtle. The Commission is proposing to redefine what counts as personal data, making it easier for AI companies to harvest information they previously could not touch. It wants to limit people's ability to access their own data. It is creating special carve-outs for AI systems that would exempt them from core GDPR protections. Companies would only need to remove personal data from AI models if doing so does not require "disproportionate efforts," a term the Commission has conveniently left undefined.

On the AI Act side, the rollback is just as aggressive. High-risk AI systems, the ones used in hiring, credit scoring, law enforcement, and immigration, would face weaker oversight. The very categories that the EU spent years debating and defining are being quietly diluted before the law has even been fully implemented.

The lobbying pressure behind this is staggering. Amazon alone spent 7 million euros on EU lobbying in a single year. Big Tech has been framing regulation as an innovation killer, and the Commission appears to have bought the argument wholesale.

Amnesty's argument is straightforward: this is not simplification. This is deregulation dressed in bureaucratic language. The beneficiaries are corporations. The losers are the 450 million people who live in the EU and thought their data was protected.

What makes this particularly galling is the timing. The EU positioned itself as the global leader in responsible AI governance. Countries across Asia, Africa, and Latin America used the AI Act as a template for their own legislation. Now the template is being hollowed out from the inside, and every country that followed Brussels' lead is left holding a blueprint for a building that the architect is quietly demolishing.

We covered the EU's deregulation push last week from the policy angle. But Amnesty's intervention elevates this from a regulatory story to a human rights story. When the organization that monitors torture, political prisoners, and genocide says your AI policy changes threaten fundamental rights, that is not routine criticism.

The EU has a choice. It can be the place that proved democratic governance can keep pace with AI. Or it can be the place that proved even the strongest regulations collapse when the lobbying budget is large enough. Right now, the smart money is on the lobbyists.