On February 19, 2020, the president of the European Commission, Ursula von der Leyen, fulfilled one of the promises she made when taking her position: to publish a White Paper on Artificial Intelligence (AI) in the first 100 days of her mandate. With this text, open for public consultation until May 19, 2020, the European Union clarifies its general outlines to face one of the greatest challenges of our time: To get the best possible results from Artificial Intelligence without leaving anyone behind or harming liberties and civil rights.
The Commission is working from the belief that the 27 can take a leadership role in the development of AI, competing with the two superpowers in this technology, China and the United States. To make up the disadvantage amassed by the EU, Brussels is looking to mobilise a public-private investment of 20 billion euros over the next decade. But it is not only a question of economics. The goal, said Von der Leyen in the presentation of the White Paper, which was created jointly with the European data strategy, is that “Digital Europe reflects the best of our continent: openness, equality, diversity, democracy and confidence”.
Two groups to combine growth with security
Beyond the general statements, the Commission’s work outline in regard to Artificial Intelligence is based on an important division: the most sensitive sectors and those with lower risk. The objective is to develop “clear standards that address AI systems having a high level of risk without creating an excessive burden for those involving lower risks”.
The first group, the sensitive sectors, includes all uses of Artificial Intelligence in the fields of health, security and transportation, such as future autonomous vehicles. In these cases, the Commission will require the use of AI to be transparent and traceable, and to include human verification. “The authorities must be able to test and certify the data used by the algorithms, as is currently done with cosmetics, passenger cars or toys,” says the Commission, concerned with ensuring “respect for fundamental rights, particularly non-discrimination”.
Nevertheless, the Commission seems to have softened their prerequisites in regard to facial recognition technologies and their cross-application with AI. After putting on the table a five-year moratorium on its use – until ethical and privacy implications in the use of these techniques is clarified – in the end the white paper only notes that use of facial recognition is only allowed under exceptional circumstances.
The requirements for use of AI in less sensitive sectors are more lax, although the Commission warns that “the strict EU regulations on consumer protection, which address unfair trade practices and protect personal data and privacy, remain applicable”.
The European Association for Digital Transition welcomes all initiatives that will allow the EU to forge its own path in the face of economic and social digitalization, abandoning the passive attitude which, unfortunately, it has been taking. It could be that this work is arriving a bit late, and might lack specificity and the necessary financial stimulus, but at least this white paper demonstrates that the crucial step has been taken: Europe has finally woken up and wants to define its own path in the development of Artificial Intelligence.