The EU AI Act: Can Regulation Make Artificial Intelligence Trustworthy?

On 21 April the European Commission published its proposal for an Artificial Intelligence Act. It is the first serious attempt by a major economy to regulate AI horizontally, across sectors, through binding law rather than voluntary principles.
The design choice at its heart is a risk based approach. Instead of regulating AI as a technology, the proposal regulates uses.
Unacceptable risk. A short list of practices to be prohibited, such as social scoring by public authorities and certain forms of manipulation.
High risk. Systems used in areas such as recruitment, credit, education, critical infrastructure, law enforcement and access to public services, which would face requirements for risk management, data quality, documentation, human oversight and conformity assessment.
Limited risk. Systems such as chatbots, subject to transparency obligations, for example telling people they are interacting with a machine.
Minimal risk. Everything else, which is most AI.
Proposed fines for the most serious violations reach up to six percent of global annual turnover.
Why it matters beyond Europe
Europe has done this before. GDPR became a de facto global standard because companies found it easier to apply one high standard everywhere than to run separate data practices by market. The Commission clearly hopes for the same effect with AI. Many companies outside Europe will design their systems to the European standard simply because they want access to the European market.
The questions for business
For European companies, the proposal turns AI ethics into a compliance agenda with very concrete questions.
Inventory. Which AI systems are we actually using? Many organizations do not know, because AI arrives embedded in purchased software.
Classification. Which of them would fall into the high risk category?
Roles. Are we a provider of the system, or a user of someone else’s? The obligations are different.
Capability. Do we have the documentation, data governance and monitoring that the high risk requirements imply?
The innovation argument
The main criticism of the proposal is predictable. Regulation will slow innovation and push AI development to the United States and China, where rules are lighter. There is something to it. Compliance costs fall hardest on smaller companies, and uncertainty discourages investment.
But I think the argument is incomplete. The largest barrier to AI adoption in regulated industries today is not regulation. It is the absence of clear rules. Banks, insurers and hospitals hesitate to deploy AI in sensitive decisions because they cannot predict how regulators and courts will judge them later. Clear rules, even demanding ones, can reduce that uncertainty.
The real risk is not regulation itself. It is regulation that is vague, slow and inconsistently enforced. That is the part to watch as the proposal moves through Parliament and Council over the next two or three years.
Leadership question: Can regulation create trust in AI without slowing down innovation and putting European companies at a competitive disadvantage?


