On March 25, the Center for Data Innovation hosted a webinar to discuss Europe’s AI priorities over the next six months. The window for the EU to establish an AI strategy to strengthen its position in the global digital economy is rapidly narrowing. Europe has many assets to build a strong AI ecosystem but trails behind its peers in terms of adoption, commercialization, and investment. In addition, EU policymakers need to establish a regulatory environment that will promote innovation and enable access to data.
According to Irina Orssich, team leader AI at the unit Technologies and Systems for Digitizing Industry at the European Commission’s DG CONNECT, the COVID-19 crisis has highlighted the importance of digital and how critical it is to society. The first priority should be to acknowledge and further explore how digital and AI in particular can contribute to the fight against the virus, such as with 3D printing of respiratory masks, or by saving scientists months of experimentation to find cures or treatments. Regarding the white paper on AI, the Commission is conducting a public consultation until May 31, and does not currently foresee an extension to this deadline. This could change depending on how the crisis evolves. The Commission will hold stakeholders meetings, public hearings, and targeted consultations with experts, including online, to assess the impact of legal provisions on various stakeholders. A large public consultation event will be organized mid-September. The Commission would still present proposals for a regulatory framework in December. In parallel, there will be an update of the coordinated action plan with member states, focusing on national AI strategies.
Elijus Čivilis, vice-minister at the Ministry of the Economy and Innovation of the Republic of Lithuania, acknowledged that the current crisis has led to AI losing priority as it is no longer on top of policymakers’ agendas. However, there is a considerable opportunity for tech and AI companies to shine: Overnight, technology has been given a huge mandate to build a new world.
Magdalena Piech, head of regulatory affairs at Allegro and chair of the European Tech Alliance, agreed that the crisis has shown how digital is essential. The crisis is putting the digital preparedness of Europe to the test and shows how critical it is. Regarding future European AI priorities, Piech referred to the role of AI in providing solutions, improving processes, and increasing efficiency. Referring to the piloting phase in which Allegro took part, Piech specified that in a vast majority of processes, in principle, there is not high risk to fundamental rights and where there is a risk, rights are protected by GDPR safeguards. Solutions need to be proportional to the risks, the regulation shouldn’t overlap with existing frameworks such as the GDPR, and new rules should be enforceable, applicable, and safeguard the competitiveness of European companies. Regarding the voluntary labeling of low-risk AI, Orssich stated that the Commission is waiting for the input of the consultation, the high-level expert group on AI, and the stakeholders who are testing the guidelines’ assessment list.
Sylwia Giepmans-Stepien, manager of public policy and government relations at Google, mentioned that AI can be used to support the workforce in areas such as content moderation. AI and supercomputers could also be used to investigate and predict the virus’ structure, as is the case with Alphabet’s AI research unit DeepMind. This can provide clues for researchers to develop vaccines. This research used non-European open source data, which shows that technology and AI call for global solutions, and that AI-training models need to use global data, not just European data. Giepmans-Stepien agreed with the general direction the Commission is taking, but recalled that the crisis shows there would be a risk and a cost of not using AI, especially in urgent situations. This is also important when defining what high risk is. Given the diversity of applications, some may present risks while others may not, and there can be no one-size-fits-all solution.
Orssich highlighted the Commission’s focus on creating the right infrastructure for sharing data such as by ensuring interoperability, contracts, and clear rules as well as by clarifying questions around competition policy. Regarding the possibility for the EU’s current strategies to be reconsidered on account of the COVID-19 crisis, the Commission will study and review the response they receive through the public consultation.
Eline Chivot, senior policy analyst at the Center for Data Innovation, raised the role of member states in implementing a common European approach to AI. Čivilis made the case that in this respect, success relies on the participation and commitment of all EU countries, through an approach that needs to be both top-down with the Commission and bottom-up with member states. It is very important for every country to work on their own understanding of the impact of AI.
Piech stated that AI is critical for European countries to be able to innovate and compete globally. She also recognized the need to identify and address some challenges that AI may pose such as legal uncertainty. On behalf of the European Tech Alliance which published its high-level principles on AI, she advocated that there does not need to be further regulation on low-risk AI applications, and that self-regulation is a potential alternative. Some other recommendations of the Alliance were the creation of a pan-European data commons, ensuring AI regulations are future-proof, fostering synergies between public and private research, and retaining the best talents in Europe.
Giepmans-Stepien argued that the global coordination of different approaches to AI is important on both investment and regulatory levels. One element she believes is not highlighted enough is the deployment of AI in Europe. Not every company needs to become AI experts, but AI can be adopted to support small businesses. Europe should focus on skills and increase the capabilities of existing businesses which are going to be the users of technology.
Čivilis said it was important not to underestimate small businesses. In Lithuania this week, after declaring quarantine, the government hosted a webinar—there is no quarantine on the Internet. Overnight, 72 SMEs had set up their ecommerce platform.
Orssich advocated for a coordinated European plan on AI. There should be a European forum to facilitate the exchange of best practices. For the Commission, there should be an attempt to seek synergies and combine the efforts of member states. She recalled that one part of the white paper focuses on a public-private partnership for AI and robotics. The idea is to ensure member states are able to do as much as possible, while trying to combine their work to get the best outcome and build on Europe’s strengths.
Regarding transatlantic priorities, Orssich added that it is key to keep the dialogue going. The Commission’s intent is not only about finding European complementarities and synergies between member states, but also transatlantic complementarities.
Matthew Glavish of the American University, D.C. contributed to this interview.