EU regulates AI: opportunities and risks?
The world's first comprehensive AI law is now a reality: the EU Parliament on Wednesday approved the version of the Artificial Intelligence Act negotiated with the EU member states. Practices such as social scoring or emotion recognition in the workplace will be banned, with exceptions for law enforcement authorities in the case of facial recognition. For some commentators the law has been too watered down. Others see it as a pioneering success.
Ahead by a nose
La Stampa praises the new rules:
“The US is clearly ahead in the technology race, even against China. But the rules are important, and Europe has won on this terrain. Because all those who are now developing artificial intelligence in the world will have to deal with these regulations: the 27 countries of the Union are too big a market to be ignored, as we have seen with the GDPR, the European Data Protection Regulation, which has become the benchmark for everyone else.”
Too many exceptions
The process that led to the AI Act shows where the real power lies within the EU, writes the Frankfurter Rundschau:
“With the governments of the individual EU states, not in the directly democratically legitimised European Parliament. Little remains in the AI Act of the Parliament's stance which aimed to defend fundamental rights against new technological means of attack - for example regarding the intended ban on facial recognition in public. The text allows too many exceptions. The fact that the EU border regime, for example, is excluded from the scope of the AI Act shows how selective the protection of human rights in the EU can be.”
Learning from past experience
These rules will promote fair competition, Corriere della Sera hopes:
“The experience with the Internet still stings. The lack of regulation has meant that the Internet market is now dominated by a handful of big players. Thanks to this new legal framework, however, a playing field is emerging on which honest, fair and equitable competition is possible. And perhaps this will facilitate the emergence of new players. Hopefully unlike the digital Wild West we've experienced to date.”
AI lobby stronger than civil society
For the taz, the business-friendly result highlights the problems of EU legislation:
“It is also the result of a negotiation situation in which business representatives are privileged over other groups, such as civil society organisations. According to the Lobbycontrol initiative, last year 78 percent of the meetings of high-ranking Commission officials on the topic of AI were with companies or business associations. The contacts are also close in vocal member states like Germany and France: in Germany, the AI start-up Aleph Alpha has good contacts right up to the ministerial level, and the situation is similar in France with AI start-up Mistral.”
Pushing key topics in the right direction
The EU must take bold action in this and other areas, La Libre Belgique insists:
“This legal framework can help Europe to re-establish itself as a major actor on the global economic stage. And at the end of the day, Europe has no other choice. High levels of investment are needed not only for the digital transformation, but also to meet the challenge of its ageing population. In other words, sources of funding must be found for healthcare and pension spending. ... Not to mention the investments that need to be made in the climate transition and defence.”