(a) Cleared from scrutiny; (b) cleared from scrutiny; drawn to the attention of the Business, Energy and Industrial Strategy Select Committee and the Digital, Culture, Media and Sport Select Committee
(a) Communication from the Commission to the European Parliament, the European Council, the Council, the European Economic and Social Committee and the Committee of the Regions: Coordinated Plan on Artificial Intelligence; (b) Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions Building Trust in Human-Centric Artificial Intelligence
Digital, Culture, Media and Sport
(40281), 15641/18 + ADD 1, COM(18) 795; (40507), 8396/19, COM(19) 168
1.1This chapter takes stock of the European Union’s emerging approach to engaging with Artificial Intelligence (AI), and the possible implications of EU exit for this policy area.
1.2On 30 April 2018 the European Commission published its “European initiative on Artificial Intelligence”, which made the case for adopting a coordinated European approach on AI in order to maximise the growth opportunities it presents, as well as to address the challenges posed by it.
1.3The strategy proposed:
1.4In the Government’s Explanatory Memorandum, the Minister for Digital (Margot James MP) noted that the initiatives proposed were similar to those recommended by the 2017 independent review produced by Professor Dame Wendy Hall and Jérôme Pesenti on how to grow the AI industry in the UK, and added that the Government would carefully scrutinise any detailed proposals in due course. The Committee cleared the document from scrutiny in its report on 24 October 2018.
1.5On 7 December 2018 the Commission published a coordinated action plan which had been developed in conjunction with the Member States and an accompanying Communication providing an overview of it. The broad thrust of the Plan is to ensure that the EU as a whole can compete globally, to improve economic growth across all sectors, to maximise the impact of investments at EU and national levels, and to encourage synergies across the policy area, including ethics.
1.6A wide range of future actions are envisioned. The most salient of these include:
1.7In the Government’s response to the Action Plan, rather than providing a subsidiarity assessment of the Commission’s proposal, the Minister simply repeated the Commission’s own assessment that “ensuring coordination on AI within Europe is necessary to remain competitive and a compelling investment opportunity at the global level” and that without it, “there is no guarantee that Member States’ individual approaches to developing AI technologies and capabilities would align and produce a compelling, optimised outcome”, without providing any assessment of this view. We interpret this as a tacit endorsement of the Commission’s analysis, as no contrary view is offered.
1.8Concerning EU exit, the Minister noted that “the future of the UK’s participation in EU regulatory regimes, policies and programmes that will impact the development and adoption of AI (such as in relation to funding, accessing data, and safety and liability frameworks) will be a matter for Brexit negotiations”, and that “the terms of participating in the Coordinated Action Plan on AI are not set out in the Communication”.
1.9The Minister also touches on the various themes of the Action Plan, although for the most part avoids commenting on what the actual policy implications of non-participation in these elements of the European aspects of the Plan might be, and instead primarily outlines the Government’s domestic agenda in the area. Thus, under “Investment”, the Minister makes no mention of the loss of funding under EU financial programmes and instead provides a summary of the UK’s AI Sector Deal.
1.10On 8 April 2019 the Commission also adopted a further Communication, ‘Building Trust in Human-Centric Artificial Intelligence’, highlighting the publication and launch of a piloting phase for the ‘Ethics Guidelines for Trustworthy AI’, developed by the Commission’s High Level Expert Group (HLEG)—a multi-stakeholder group representative of different Member States alongside academic and industry perspectives. The piloting phase will be launched by the summer of 2019 and last until the end of the year. After this piloting phase they will be further revised at the beginning of 2020.
1.11The guidelines highlight trust as a prerequisite to ensure a human-centric approach to the development of AI. The document captures the founding principles of the EU as guiding principles for this work (respect for human dignity, freedom, democracy, equality, the rule of law and respect for human rights), as well as the EU Charter of Fundamental Rights which captures the personal, civic, political, economic and social rights enjoyed by people within the EU. The seven key requirements relate to: human agency and oversight; technical robustness; privacy and data governance; transparency; diversity, non-discrimination and fairness; societal and environmental well-being; and accountability.
1.12The Minister states that the Government considers the non-binding guidelines to be “a helpful contribution to the global debate on the ethical deployment and governance of artificial intelligence”. She also welcomes “the rights-based approach proposed in the guidelines as a more practical way to define the ethical challenges surrounding the deployment of AI”, and notes that the guidance incorporates the UK’s feedback and complements the UK’s Data Ethics Framework, which sets out clear principles for how data should be used in the public sector.
1.13We have taken note of the Minister’s views regarding these communications respectively setting out a coordinated action plan and ethical guidelines regarding Europe’s approach to Artificial Intelligence (AI).
1.14In terms of the implications of EU exit for AI in the UK, the Minister states that “the future of the UK’s participation in EU regulatory regimes, policies and programmes that will impact the development and adoption of AI (such as in relation to funding, accessing data, and safety and liability frameworks) will be a matter for Brexit negotiations”.
1.15Our assessment is that the primary questions that arise for AI stakeholders in the UK as a consequence of Brexit concern, not the emergent EU AI regime, which currently foresees a cooperative approach which could accommodate a degree of divergence and experimentation at Member State level, but to (i) EU rules relating to data, (ii) the extent of UK participation in EU financial programmes, and (iii) access to talent.
1.16In terms of data, once the UK has left the EU (and any transition period has come to an end), it will cease to participate directly in the many EU frameworks governing data (including personal data protection, the free flow of non-personal data, and the re-use of public sector information) which cumulatively have the effect of facilitating access to and use of data throughout the Union, which will instead apply to the UK as a third country. As access to large datasets are integral to AI, any increased friction in this regard could negatively impact UK stakeholders. The Minister’s acknowledgement that “the European ambition to create a common data space in discrete sectors or areas recognises the need for large amounts of high-quality data to develop AI technologies” implicitly acknowledges these concerns, highlighting in particular the importance of the High Value Datasets which may be created through Implementing Acts under the recast PSI Directive. In the event of a non-negotiated withdrawal, the Commission has stated that UK-based stakeholders would have to use the tools under the General Data Protection Regulation to meet their needs in a ‘no-deal’ scenario, which (until an Adequacy Decision or other bilateral arrangement between the UK and the EU were concluded) would require businesses to incur a range of costs to continue to collect, process and transfer the personal data of Union citizens.
1.17Leaving the Union will also result in the loss to UK stakeholders of substantial EU funds for AI-related research unless the UK negotiates association to/participation in the relevant financial programmes. If the Multiannual Financial Framework (MFF) proposed by the Commission were to be approved, this would mean that at least €1bn per year would be made available to be invested in AI during the period 2021—2027 under the Horizon Europe and Digital Europe programmes, in addition to EU investments in High Performance Computing infrastructure, AI startups, testing and experimentation sites for AI-powered products and services, and for the promotion of advanced AI skills. UK participation in the Horizon Europe and Digital Europe programmes would have to be negotiated with the EU within the context of wider negotiations relating to the overall relationship. We note that the proposed terms governing the participation of third countries in the Horizon Europe Regulation require the Union “to take into account the objective of driving economic growth in the Union through innovation” and states that “parts of the Programme may be excluded from an association agreement for a specific country”—a provision which could potentially be used to limit the involvement of third countries in AI-related elements of the programme. In the event of a non-negotiated exit, UK participation in the current Horizon 2020 programme would be disrupted as there would be no implementation period, and it would be politically challenging to negotiate association to Horizon Europe until wider exit issues had been addressed.
1.18Other issues affecting AI stakeholders in the UK as a result of EU exit include:
1.19We now clear these two non-legislative documents from scrutiny. We will scrutinise the development of the AI Action Plan through subsequent documents. We draw this report to the attention of the Select Committee for Digital, Culture, Media and Sport and the Select Committee for Business, Energy and Industrial Strategy.
(a) Communication from the Commission to the European Parliament, the European Council, the Council, the European Economic and Social Committee and the Committee of the Regions: Coordinated Plan on Artificial Intelligence: (40281), 15641/18 + ADD 1, COM(18) 795. (b) Communication from the Commission to the European Parliament, the Council, the European Economic and Social Committee and the Committee of the Regions: Building Trust in Human-Centric Artificial Intelligence: (40507), 8396/19, COM(19) 168.
None, but see: Forty-first Report HC 301–xl (2017–19),(24 October 2018).
1 Communication from the Commission to the European Parliament, the European Council, the Council, the European Economic and Social Committee and the Committee of the Regions—Artificial Intelligence for Europe, .
2 Explanatory Memorandum from the Minister to the Chair of the European Scrutiny Committee ().
3 Growing the Artificial Intelligence Industry in the UK ().
4 Forty-first Report HC 301–xl (2017–19), (24 October 2018).
5 European Commission, Coordinated Action Plan on Artificial Intelligence, .
6 European Commission, Annex to the Coordinated Plan on Artificial Intelligence, .
7 Explanatory Memorandum submitted by the Department for Digital, Culture, Media and Sport ().
8 HM Government, Policy paper: AI Sector Deal ( ).
9 European Commission, Communication: Building Trust in Human-Centric Artificial Intelligence, .
10 European Commission, High-Level Expert Group on Artificial Intelligence: Ethics Guidelines for Trustworthy AI ().
11 Explanatory Memorandum submitted by the Department for Digital, Culture, Media and Sport ().
12 Explanatory Memorandum submitted by the Department for Digital, Culture, Media and Sport ().
13 European Commission, Proposal for a Directive on the re-use of public sector information, .
14 The 13 November 2018 European Commission Communication, “” states that “In the case of a no deal scenario, as of the withdrawal date, the transfer of personal data to the United Kingdom will become subject to the rules on international transfers in application of the General Data Protection Regulation (EU) 2016/679, Directive (EU) 2016/680 for the law enforcement sector and Regulation (EC) 45/2001 26 for the institutions and bodies of the European Union. The General Data Protection Regulation, Directive 2016/680 and Regulation 45/2001 contain a broad toolbox for data transfers to third countries. This includes in particular the so-called ‘appropriate safeguards’ (e.g. the Commission’s approved Standard Contractual Clauses, Binding Corporate Rules, administrative arrangements) that can be used both by the private sector and public authorities.”
15 European Commission, Proposal for a REGULATION OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL establishing Horizon Europe—the Framework Programme for Research and Innovation, laying down its rules for participation and dissemination, (Article 12(2)).
16 TechUK, An immigration system that supports the UK tech sector ().
Published: 2 July 2019