The challenges posed by Artificial Intelligence (AI) are numerous. They are linked in particular to problems of explicability of algorithms, opacity of operation and complexity. In order to regulate them, a horizontal European regulation was necessary. This new regulation, the draft of which was published on 21 April 2021, is intended as a strong political act.
The new rules applicable to AI follow a risk-based approach
- Unacceptable risk (going against European values): AI systems considered a clear threat to people’s safety, livelihoods and rights will be banned. This includes AI systems or applications that manipulate human behaviour to deprive users of their free will (e.g. toys using voice assistance to induce minors to engage in dangerous behaviour) and systems that enable social rating by states.
- High risk: AI systems that will be subject to strict evaluation rules, including AI technologies used in:
- critical infrastructure (e.g. transport) that could endanger the lives and health of citizens;
- education or vocational training, which may determine a person’s access to education and career path (e.g. marking of exams);
product safety components (e.g. application of AI in robot-assisted surgery);
- employment, workforce management and access to self-employment (e.g. CV sorting software for recruitment procedures);
- essential private and public services (e.g. credit risk assessment, which deprives some citizens of the possibility to obtain a loan);
- the area of law enforcement, which may interfere with the fundamental rights of individuals (e.g. checking the reliability of evidence);
- the area of migration management, asylum and border control (e.g. checking the authenticity of travel documents);
- the areas of administration of justice and democratic processes (e.g. application of the law to a concrete set of facts).
- Limited risk: AI systems to which specific transparency obligations apply.
- Minimal risk: the legislative proposal allows for the free use of applications such as video games or spam filters based on AI. The vast majority of AI systems fall into this category. The draft regulation does not provide for intervention in this area, as these systems pose little or no risk to the rights or safety of citizens.
Finally, it should be noted that some AI-specific issues will also be introduced in some sectoral regulations, such as the introduction of AI-related requirements in the draft revision of the Machinery Directive.