European Commission unveils AI regulations

BRUSSELS, April 21 (Xinhua) -- The European Commission presented its first ever legal framework on artificial intelligence (AI) on Wednesday.

As part of the AI Package adopted by the executive of the European Union (EU), the AI legal framework aims to build trust into AI for the EU's population and ensure excellence, according to Margrethe Vestager, European Commission's executive vice-president responsible for competition.

According to their level of risk, the European Commission divided different types of AI applications into four categories: minimal, limited, high and unacceptable.

"The higher the risk that a specific use of AI may cause to our lives, the stricter the rule," explained Vestager during her presentation.

Minimal risk AI systems can be developed and used without any restrictions. These are the vast majority of currently used AI systems in the EU, like spam recognition software.

Limited risk AI systems will be subjected to transparency requirements to make sure that the user knows he interacts with a machine and not a human being, like the case with chatbots.

High risk AI systems are those which can interfere in a person's life in a significant way. These systems will be subjected to a series of strict obligations, including adequate risk assessment and mitigation system, high quality of data set, detailed documentation, appropriate human oversight, among others.

High risk AI systems cover critical infrastructures, educational or vocational training, safety components of products, employment, workers management and access to self-employment, and so on.

All those AI systems that are "considered a clear threat to the safety, livelihoods and rights of people" are unacceptable and thus banned. These include, for instance, toys encouraging dangerous behavior in minors through voice assistance.

Sanctions for manufacturers or users that do not comply with the requirements will first be asked to do so. Later on, the product might be withdrawn from the market. In case of repeated offenses, fines of six percent of their yearly global turnover will be applied.

The AI package presented by the European Commission also includes the New Machinery Regulation and an updated Coordinated Plan on AI.