DIGITALEUROPE welcomes the Artificial Intelligence Communication but warns more consistency on policy is needed
The EU has an opportunity to be a world leader in AI but it can be undermined by its own proposals for Copyright reform, ePrivacy Regulation, Consumers Rules and the Platform-to-Business Regulation.
Today the European Commission announced its Communication on ‘Artificial intelligence for Europe’. It will aim to address an open dialogue on ethics as well as accelerate the adoption of AI technologies by businesses. We welcome this initiative and will be ready to support the European Commission in these efforts. However, only 10-35% of large to small enterprises are using big data analytics. More needs to be done to accelerate adoption.
In addition, the Communication addresses a lack of skills and investments in the EU to be able to keep pace with regions such as China and the US. Addressing these issues will help the EU reap the benefits of AI for economic growth and to help deliver solutions to important social issues.
Unfortunately, various policy proposals, currently in the legislative process, contain rules that negatively impact the development and use of algorithms and if they were approved today, it would slow down the benefits of AI technologies for European businesses and ultimately impose barriers for Europe to compete internationally.
“The AI Communication plans to address severe issues such as the lack of investment and digital skills. Thecurrent prediction is that 83% of all AI investment will be done outside Europe1. If we do not urgently tackle this head-on, we risk diminishing the EU’s role and relevance on the international stage and hampering jobgrowth in Europe.
It’s also of utmost importance to discuss the implication of AI on the society and jobs, whereby we shouldfirst assess, if or, where existing regulation is already in place or requires to be updated. Let’s make sure thatwe embrace AI in society and foster innovation, by building up digital skills for both the young and old.
We unfortunately see several diverging policy proposals, which legislators and stakeholders need to addresswhile there’s still time. It’s more than ever the case for a consolidated and comprehensive European AI strategy.”
DIGITALEUROPE encourages that all ongoing and future legislative proposals undergo an additional review with regards to their impact on AI to be developed; in an open stakeholder dialogue. These discussions should also take place in advance to any single legislative files aiming to regulate algorithms unilaterally through restrictions, requirements and fines. DIGITALEUROPE’s key concerns are with the following proposals:
The Consumer Rights Directive (art. 6a) could require additional information for contracts to consumers to include the main parameters on the ranking of offers that result from a search query on an online marketplace.
In the Copyright Directive, the provision on text-to-datamining (art. 3) would mean that the large community of research partnerships, SMEs and tech companies would obliged to pay additional fees and require extra permissions for R&D. The text on upload filters (art. 13) would further only exempt non-commercial and private services form the new liability rules and hence catch all business cloud services for AI delivery.
The Regulation ‘on promoting fairness and transparency for business users of online intermediation services and online search engines in the Digital Single Market’ (P2B Regulation) has the potential to be amended by legislators looking to impose rules on the complex area of ranking by algorithms.
The proposed ePrivacy Regulation would greatly hamper the development of assistive communication features across all sectors (connected cars, health care) as they are based on machine learning.
DIGITALEUROPE wants to engage policymakers to be able to address all barriers preventing the adoption and development of AI. This includes exchanges on how AI will be used and ethics around these uses, the quality of data being processed, and the transparency and predictability of algorithms as well as the explainability of the outcomes. Each of our members involved in the development of AI are taking these issues seriously and instituting rules and procedures in their engineering processes.
Our members and national trade associations stand ready to discuss this topic with the co- legislators. We invite the European Commission, the Parliament and the Council of the EU to maintain an open dialogue with DIGITALEUROPE.