European Union is proposing the first law to regulate Artificial Intelligence
After several strategies adopted in the field of artificial intelligence (AI) and a consultation with the market, the community executive will, on Wednesday next week, present a proposal for regulation to safeguard the fundamental values and rights of the EU and the user safety, aiming to compel systems considered as high risk to comply with mandatory requirements related to their reliability.
European Union is proposing the first law to Artificial Intelligence
In this dossier that was taken over as a priority by the presidency of the EU, Brussels will present the first legislation aimed at AI, although the creators and those responsible for the implementation of this technology are already subject to European legislation on fundamental rights – for example, data protection, privacy, non-discrimination – consumer protection and product safety and liability rules.
AI has been increasingly used in areas such as entertainment (personalization of content), online commerce (forecasting consumer tastes), home appliances (intelligent programming) and electronic equipment (using virtual assistants as Siri or Alexa, among others).
The European Commission has tried to strengthen cooperation between Member States regarding AI, but there is still no common legal framework, which is why the institution is now taking steps towards standardization, aiming at the development of technology, used in areas such as cars, cosmetics or toys.
The aim is to move from a voluntary approach to the regulatory sphere, which will still not be too rigid given the constant evolution of technology, but rather flexible and interconnected with existing European rules, according to the sources.
“We don’t want to demonize AI or say, ‘Oh my God, what a terrible technology.’ We want to create a trusted ecosystem to ensure that the risks associated with this technology are properly regulated,” says the European sources.
Therefore, a list of criteria will be created to verify that a given system is not of high risk and, for those that are really seen as risky, there will be legal requirements such as the obligation to provide information and human supervision, as well as rules on conservation of records and data.