On 19 February 2020, the European Commission published two important documents: White Paper on Artificial Intelligence: A European approach to excellence and trust and A European strategy for data.
The White paper emphasizes the need to build trust and to ground the European AI in European values and human rights. In order to achieve secure and trustworthy AI, the EU should mainly build an ecosystem of excellence and an ecosystem of trust.
With regard to the ecosystem of excellence, the White paper proposes a number of actions: working with Member States on revision of the Coordinated Plan on Artificial Intelligence, creation of new centers of excellence and centers for testing that could combine EU, national and private investments, developing necessary skills by establishing networks of leading universities and attracting the best professionals in the field, supporting specialization of national digital innovation hubs in specialization in AI and financing innovative developments in AI, setting up new public private partnerships in AI, promoting use of AI by the public sector, “securing access to data and computing infrastructures”, and promoting EU values on the international level.
With regard to the ecosystem of trust, the Commission highlights the need to efficiently secure protection of fundamental human rights and legal certainty. The Commission refers to Guidelines on trustworthy AI and its seven key requirements. At the same time the Commission defines two types of risks connected with AI: risks for fundamental human rights and risks for safety and determining liability. The Commission suggests to improve the legislative framework and at the same time states that the new regulatory framework should not be excessively prescriptive. For that purpose, the framework “should follow a risk-based approach”. The White paper defines criteria for determining what a high-risk AI application is. Examples of high-risks applications are AI applications in recruitment process or remote biometric identification. The White paper also formulates types of future requirements in line with Guidelines on trustworthy AI and suggests prior conformity assessment of AI applications.
The White paper shall be open for consultation until 19 May 2020.
The communication A European strategy for data emphasizes the importance of data in the data economy. The Commission identifies a number of problems that need to be addressed: availability of data, imbalance in market power, data interoperability and quality, data governance, data infrastructures and technologies, empowering individuals to exercise their rights, skills and data literacy, and cybersecurity. In the remainder of the document, the Commission formulates action steps in order to solve the aforementioned issues.
The Annex of the document provides an overview of specific sectors important for creation of data spaces.