AI in Europe - Coordinated Plan in a Nutshell

Last Friday, the European Commission published a "Coordinated Plan for Artificial Intelligence", as we committed to do when we published our AI strategy back in April. Since then, we held regular discussions with Member States, and the result is the new Coordinated Plan.

In a nutshell, the Plan lists actions on AI that the Commission and Member States plan to launch in the immediate future, and in doing so it ensures synergies and avoids wasteful duplication - hence its name “Coordinated Plan”. The plan is based on annual reviews and updates to keep up with developments.

The approach set out in the plan follows the logic of the AI Strategy: it focuses on investment to ensure Europe’s presence at the forefront of AI, it addresses societal issues such as the impact on skills requirements, and it underlines the urgent need for an ethical dimension to be taken into account from the start. In addition, it also looks at availability of data and the use of AI in public services.

As far as Europe’s AI capacity is concerned, the Plan encourages all Member States to adopt AI strategies by mid-2019. It also suggests that Member States match the recent increase in AI funding in the EU’s research programme – by 70% to EUR 1.5 bn – so that overall investment in AI reaches EUR 20 bn over three years, rising later – in the course of the next decade - to EUR 20 bn annually. To better direct this funding into productive investment, the plan proposes improved partnerships between the European institutions, Member States, research establishments and industry, to strengthen research and innovation as well as capacity in testing and experimentation facilities. It also engages digital innovation hubs to diffuse AI to SMEs and to the public administration.

On the rising skills requirements, the plan notes that a triple focus is needed to train, retain and attract AI specialists. Indeed, while more training is certainly needed, it is also becoming increasingly difficult to retain specialists, since they receive lucrative employment opportunities in third countries. Moreover, Europe's ability to attract foreign cutting-edge researchers could be improved.

Regarding the availability of data, the European Commission and the Member States commit to improved reusability of high-value public data sets and to strengthening the development of industrial data platforms. We also look at the opportunities of enhancing data platforms in healthcare (through data that are anonymised and based on donorship), so that AI can be trained to help improve diagnoses and treatments.

On the ethical dimension, the concept of “ethics by design”, which requires ethical principles to be embedded in AI products and services right at the beginning of the design process, is the key element of Europe's approach. Integrating ethical considerations into the technological development will be a specificity of the European approach and can even become a competitive advantage over time, as the need for an ethical approach will be more widely recognised.

In this context let me recall the High-Level Expert Group on AI, which we had established in order to draft some widely acceptable ethical guidelines. After the finalisation of the first draft, there will be a one-month period for comments, so please feel free to make your point of view known.