Setting a European vision for trustworthy AI

In recent times, Artificial Intelligence (AI) has grown to become an area of strategic importance and a key driver of economic development. Alongside other international actors, Europe has taken its first steps to become a global hub for trustworthy AI, while sticking to its foundational rights and values. Michał Zakrzewski, APPLiA Digital & Competitiveness Policy Director, was invited to participate in the discussion on the EU’s proposed Artificial Intelligence Act, on behalf of the home appliance industry, at the European Forum for Manufacturing. 

Whilst it is key for the EU not to miss this global race from a technological, economic, and geopolitical point of view, “it is of critical importance to work towards the development of a better understanding of the matter,” kicked off Zakrzewski.  For a long time, European consumers have had their grasp of AI influenced by sci-fi novels and movies which do not reflect the reality for most European manufacturers. For this reason, “a common definition of AI is needed to provide legal certainty to manufacturers and regulators, alike.” A prerogative that appears to be addressed by the draft Artificial Intelligence Act, with a need for improvements over the negotiations with co-legislators. As the first ever attempt made by the EU to enact an horizontal regulation on the matter, the Act paves the way for a wide range of opportunities both for European businesses and citizens. 

As such, Zakrzewski underlined that “flexibility should be the master ingredient of a successful AI recipe”, that is functional to the needs of EU industries. “Over-regulation would indeed risk creating too many barriers to technological innovation while ultimately negatively impacting the competitiveness of the industry,” he detailed. With this regard, it is of critical importance that freedom to research and innovate is preserved at all stages. 

While the potential of AI for citizens, businesses and public interest is clear, preserving consumer safety remains a key prerogative for the industry. On this note, Zakrzewski pointed to the importance of “striking a fair balance between safety and innovation, with an eye at strengthening Europe’s ability to compete globally.” Let’s take the case of home appliances, in reference to Art.6 of the draft Act. Products should only be considered ‘high-risk’ if the AI system is clearly intended to be used as a core safety component or is the product itself. If, instead, the AI system is only providing additional functionalities, the product should not be considered ‘high-risk’. In this sense, the implementation of agreed standards is crucial to the successful operation of legal requirements, from principles to practice. 

The sector highly values the proposed legislation and is committed to contributing to strengthen Europe’s excellence in AI research and innovation. For this to happen, a smooth functioning of the standard-making system shall be secured. On this crucial point, Zakrzewski referred to Art.41 of the draft Act, expressing the need to amend it in a way to ensure that the role of EU Standardisation Organisations is preserved at all stages. Clear conditions shall be set on the possibility to issue common specifications via implementing acts, in order to best secure the well established standard making process, operating in the EU. 

Overall, the draft Act sets a thoughtful start to the legislative process in Europe by laying the foundations for trans-Atlantic cooperation, in the pursuit of trustworthy innovation and towards a successful capitalisation of AI in Europe.