EU Strategy on Artificial Intelligence In 2018
A Closer Look at the Strategy by the European Union
Most may at the time of writing associate EU with Brexit since the United Kingdom is pulling out of the union. The European Union and their member countries does together have a population of approximately 500 million and about $22.0 trillion GDP which places EU as the 2nd largest economic force in the world. Therefore by some measures it is an important area to keep track of, and the international strategy for EU relating to AI may be of interest. By summarising some of these policies in a pragmatic way I hope you as a reader understand that this is no substitute for reading the documents, rather an attempt to bring together a few key points. What I provide is of course not a complete picture, rather small excerpts from an ongoing discussion.
Looking at the EU strategy it can be hard to understand where to start. One can look at several different documents, and there are often new developments in the policy that is being made related to AI.
You would think this was impossible to navigate, however EU has been kind to attempt pulling together a few threads relating to this topic on their page for the Digital Single-Market. There is of course more legislation and strategy that relates to artificial intelligence, yet I am afraid we will keep to this narrow frame:
- Declaration of cooperation on Artificial Intelligence (2018, April)
- Artificial Intelligence for Europe (2018, April)
- Ethics Guidelines for Trustworthy AI (2019, April)
- Policy and investment recommendations for trustworthy Artificial Intelligence (2019, April)
- The European Alliance Assembly (2019, June)
- Liability for Artificial Intelligence (2019, November)
I will start with the two documents outlined in bold, and I may add new documents as well if I deem it necessary in the coming days.
1. Declaration of Cooperation on Artificial Intelligence
25 European countries signed a Declaration of cooperation on Artificial Intelligence on the 10th of April in 2018. Norway was part of this, although they are not an EU member (rather part of the EEA — European Economic Area). This was said to build on the pre-existing investment and community in Europe.
It set out boosting technology and industrial capacity through access to public sector data. Addressing socio-economic changes, especially in the labour market. Ensuring an adequate legal and ethical framework building on fundamental rights and values, as well as transparency and accountability.
The commitment that “Member States agree to” (I have simplified at times):
- Work towards a comprehensive and integrated European approach on AI to increase the EU’s competitiveness, attractiveness and excellence in R&D in AI (where needed modernise national policies).
- Encourage discussions with stakeholders on AI and support the development of a broad and diverse community of stakeholders in a European AI Alliance to build awareness and foster the development of AI that maximises benefit to economy and society.
- Consider allocation of R&D&I funding to the further development and deployment of AI, including on disruptive innovation and applications, as a matter of priority.
- Reinforcing AI research centres and supporting their pan-European dimension.
- Establishment of Digital Innovation Hubs at European level.
- Make AI available in public sector. Exchange best practices on procuring and using AI in government.
- Help SMEs and companies from non-technological sectors get access to AI.
- Exchange views on ethical and legal frameworks related to AI.
- Ensure humans remain at the centre of the development, deployment and decision-making of AI, prevent the harmful creation and use of AI applications.
- Advance the public understanding of AI.
- Engage in a continuous dialogue with the Commission on AI.
2. Artificial Intelligence for Europe
This communication has a more positive slant in terms of what AI can solve and gives an introduction to what AI is. The message is that AI is transforming society like the steam engine or electricity. It says a solid European framework is required.
This report says that EU should have a ‘coordinated approach’ towards AI for good and for all. It suggest to do so through:
- World-class researchers, labs and startups
- Digital Single Market — common rules for data protection, cybersecurity and connectivity
- Unlocking data (termed as ‘the raw material for AI’ by the report)
It mentions the commitment in the declaration. In almost the same breath it puts particular emphasis on competition, leaving no one behind and EU’s sustainable approach to technologies. An approach that benefits people and society as a whole.
It goes back to the review of the Digital Single Market strategy in 2017 where the invitation to explore the EU approach to AI partly originated (with civil law rules to robotics for example). The three points there was on (1) boosting tech capacity, (2) preparing for socio-economic changes and (3) ensuring ethical and legal framework.
A different part of the communication on EU’s position in the competitive international landscape outlines the increased investment in unclassified AI research by US and China, thus making it clear that EU is ‘behind in private investments’. There was an expressed wish to create an environment that stimulates investments. EU apparently produces one fourth of the professional service robots. “Europe cannot miss the train” and the benefits of adopting AI are widely recognised — a few projects funded by EU are mentioned (agri, healthcare, infrastructure and manufacturing).
In outlining the way forward they say that a joint effort by private and public is needed by 2020 and beyond. It outlines an increased investment from around EUR 4–5 billion towards EUR 20 billion over the following decade.
This way stepping up investments is stated. They outline a growth towards this number by the end of 2020. This report also mentions supporting excellence centres (and digital innovation hubs). A mention of the ‘AI-on-demand platform’ is made here that can help facilitate collaboration between the more than 400 digital innovation hubs. It mentions that hubs focused on AI will be created.
Toward 2020 they will invest 1.5 billion in research and innovation; strengthening excellence centres; and a toolbox for potential users. They talk of an AI-on-demand platform and industrial data platforms giving access to quality datasets. Beyond 2020 they will be upgrading and supporting public interest applications, and a support-centre for data sharing alongside a variety of upgrades to existing policies.
There is an aim to make more data available, EU has been doing so over the last 15 years, an examples is EU’s space programmes. It talks of an ageing society and enhancing people, in their ‘leaving no one behind’ they talk of new skills. This was largely oriented towards STEM while in another section they talk of diversity as well as interdisciplinary approaches.
- More women and people of diverse backgrounds including people with disabilities.
- Interdisciplinary ways combining joint degrees for example in law or psychology and AI. The importance of ethics is mentioned here, while considering the creation of an attractive environment to make talent stay in Europe.
The report had plans for education policies in 2018 with (re-)training schemes, analysis of labour market, digital traineeships in advanced digital skills, business-education-partnerships and social partners to include AI in impact studies.
“Proposals under the next EU multiannual financial framework (2021-2027) will include strengthened support for the acquisition of advanced digital skills including AI-specific expertise.”
In ensuring the ethical framework there is a mention of fundamental rights, GDPR, digital single market and explainable AI systems. In the last paragraph there is additionally a question of intellectual property rights. Draft ethics guidelines were to be developed towards the end of the year. Safety and liability was mentioned leading into the empowerment of individuals through guidance document on the product liability directive in light of technological development by mid-2019. A pilot project in Algorithmic Awareness Building was mentioned too and support for consumer organisations for data protection.
Engaging member states is important in this work and this section mentioned Finland’s national strategy. It says: “Every member state is encouraged to have an AI strategy, including on investment.” A multi-stakeholder European AI Alliance was mentioned here as well as international outreach.
“With AI being easily tradeable across borders, only global solutions will be sustainable in this domain.”
The contribution by EU is mentioned with its values and fundamental rights, and this is mentioned in the conclusion as well. Ending with the wish to place the power of AI at the service of human progress.
This is #500daysofAI and you are reading article 202. I am writing one new article about or related to artificial intelligence every day for 500 days. My current focus for 100 days 200–300 is national and international strategies for artificial intelligence.
