With its new ‘Shaping Europe’s Digital Future’ Strategy, the European Data Strategy and a White Paper on Artificial Intelligence, all published on the same day (19 February 2020), the European Commission led by Ursula von der Leyen is fully committed to digitalizing our society. Zeroing in on the AI White Paper, it is clear that the Commission tries to find a delicate balance between building both an ecosystem of excellence that supports the development and uptake of AI and an ecosystem of trust where AI is also regulated and safe for everyone. The European Commission has already undertaken quite some work in defining its approach to AI and in consulting stakeholders. It is now proceeding with an official written consultation, seeking feedback on the White paper through a questionnaire.
While the European Commission has the prerogative to initiate the above ideas and strategies, the European Parliament has not stood still in the past couple of months and has proactively, and extensively, positioned itself and defined its priorities. Most notably, the Parliament’s Legal Affairs Committee (JURI) is working on multiple AI reports, focused on the technology’s ethical aspect, its civil liability regime and intellectual property rights for the development of AI technologies. Furthermore, also the Parliament’s Culture and Education Committee (CULT) is working on its own report on AI applications in education, culture and the audiovisual sector, and a EP Resolution has been drafted on automated decision-making processes and ensuring consumer protection and free movement of goods and services.
While the discussion around AI at EU level seemed to have stemmed exclusively from the new Commission’s strong will to act on this issue, since then feedbacks from civil society, NGOs, companies and others, have been highly requested to shape further the future framework.
The latest opportunity for stakeholders to contribute to the discussion is the Commission’s Consultation on the White Paper on Artificial Intelligence, closing on 14 June.
The questionnaire explores certain aspects of the White Paper, including specific actions to build an ecosystem of excellence, options for a regulatory framework for AI and further consultation on the question of safety and liability aspect of AI.
To zoom in on a specific and rather important aspect of the questionnaire for businesses, the Commission is seeking feedback on whether the introduction of new compulsory requirements should be limited to high-risk applications, and whether the current definition and criteria for this risk-based approach is the right way forward. New requirements and standards would regulate aspects such as training data, human oversight and so on. In addition, the European Commission is seeking feedback voluntary labelling for any other AI-powered services that could be qualified as “low-risk”. The intention and content of such voluntary labelling scheme is still fully open for discussion.
There are two key opportunities for businesses here through this process, that should not be overlooked:
First, this should be seen as the perfect opportunity to question, understand, assess and if necessary, improve companies’ practices when developing or using AI in their daily activities. The Commission and Expert Groups have developed various tools such as the White Paper, but also the assessment list of the Ethics Guidelines for Trustworthy AI., that can guide this type of exercise. Do we allow for human oversight? Does the data we use could lead to biased decisions? Would we benefit for a voluntary label or other form of self-regulation? Those are some of the questions that companies operating in the EU could ask themselves to stay relevant in the market.
Second, share companies should share their experience with policymakers to ensure that a new EU legal framework does not hinder business activities or innovation beyond what is necessary to protect consumer and fundamental rights, and to ensure that any new legal framework does not create legal uncertainty or unnecessary red tape. Referring once again to the risk-based approached, the possible evolution of the qualification and criteria for “high-risk” use can have a significant impact on companies. Stakeholders have an opportunity to shape rules that could ensure the EU remains an open, competitive, and innovative market.
There have been certain voices calling for a reassessment of the Commission’s plans in relation to AI under the new circumstances created by the COVID19 outbreak, which could shed new light on the costs of not using AI-powered solutions. The Commission has however clearly insisted on the fact the questionnaire would be the perfect opportunity to reflect further on what a future regulation should look like to ensure that AI fulfill its promises for society.