Europe’s regulatory retreat on AI: a free lunch for Big Tech?

The noise surrounding 'competitiveness' in Brussels is reaching a fever pitch.

A recent casualty of this push for 'simplification' is the AI Liability Directive (AILD) — a legislative proposal that would introduce clear rules on accountability when AI systems cause harm.

Its conspicuous absence from the EU Commission’s 2025 work programme leaves a worrying gap in Europe's AI framework.

At the AI Action Summit earlier this year, EU leaders signalled a shift of priorities, away from safety and towards “competitiveness.”

Commission president Ursula von der Leyen declared that “we have to cut red tape” to “make it easier” for AI to grow in Europe. Just days later, the commission translated rhetoric into action and effectively withdrew AILD.

Unlike the AI Act, which focuses on reducing damage caused by high-risk AI systems, AILD was designed to ensure accountability when harm does occur and provide a clear route to compensation for those affected.

The commission’s decision to drop the file, despite its obvious importance for consumer protection, looks less like a technical decision, and more like a political concession.

The directive would have introduced legal liabilities for major AI developers, a prospect Big Tech fiercely resisted. Ultimately, tech giants do not want to be held accountable for the products they develop, and neither do the firms using them.

But someone needs to be accountable when things go wrong.

If an AI system denies credit, triggers a market crash, or locks a vulnerable consumer out of basic services, who is responsible?

As a recent Finance Watch report makes clear, financial regulation is built on the principles of accountability, responsibility and transparency.

Until recently, supervisors could identify the source of a trading error or an unjustified denial of insurance. Even the most complex software-driven decisions were ultimately explainable.

Faulty code could be identified, and responsibility assigned. This chain of accountability is embedded in financial regulation, and firms are obliged to comply. If they fail to do so, they face consequences. 

But AI breaks this chain. If an AI system denies credit, triggers a market crash, or locks a vulnerable consumer out of basic services, who is responsible?

AI Deep learning systems operate by detecting correlations in vast datasets, not through a transparent logic of cause and effect.

These models function in a ‘take it or leave it’ manner, where even their own developers struggle to explain outputs.

This 'black-box logic' renders effective oversight impractical, if not impossible. Credit assessments, insurance pricing, or investment decisions made by AI resist explanation, and regulators may struggle to detect errors, biases, or even systemic risks.

AI is not business as usual. It challenges the application of financial regulation’s core principles. And when the commission backs away from liability, a concerning gap opens in the regulatory framework. 

A free market without accountability is simply a license to exploit.

AI is already an oligopoly dominated by a handful of US enterprises, and the EU’s retreat hands them even more power. By abandoning the AI liability regime, the commission is effectively telling these companies they can benefit from the single market without bearing any responsibility for the harm their systems may cause.

EU policymakers should not get swept up in a competitiveness frenzy that confuses sound regulation with red tape. They must step back and take stock.

What’s needed now is not deregulation, but a reassessment of the AI rulebook. Pragmatic steps can be taken to ensure that, as AI use cases in finance proliferate, citizens are protected from harmful practices.