European Union: AI responsibility in context

Themes that matter for professional AI use

EU AI Act

The EU regulation on artificial intelligence introduces obligations that depend on the role of the actor and the risk category of the system. Which obligations apply to a professional user of an AI-assisted tool depends on the specific deployment.

AI literacy

Organisations are expected to ensure that people working with AI systems have an adequate level of understanding of how those systems work and where they fail.

Transparency duties

In several contexts, people must be able to recognise that content or a decision basis was produced with AI assistance.

High-risk contexts

Certain application areas carry additional requirements, including risk management, data governance, documentation and oversight.

Human oversight

Oversight is only meaningful when the person overseeing can see what the system did and can act on it. Interfaces that present output without its basis work against this.

Data protection

Processing personal data with AI assistance stays subject to the GDPR, including the legal basis, the information duties and the rules on processors and international transfers.

Documentation and accountability

Being able to demonstrate how a result came about is a recurring theme across European rules, independent of the specific instrument.

Enforcement

European instruments generally provide for supervisory enforcement and sanctions. The relevant authority and the scale depend on the instrument and the member state.