Coordinated Plan on Artificial Intelligence published 07122019 Maximise
Coordinated Plan on Artificial Intelligence (published 07/12/2019) Maximise investments through partnerships Create European data spaces Nurture talent, skills and life-long learning Develop ethical and trustworthy AI
Artificial Intelligence for Europe (published 25/04/2018) STRATEGY FOR EUROPE TO LEAD THE WAY Boosting technological and industrial capacity – AI uptake Prepare for socioeconomic changes Ensure an appropriate ethical and legal framework Development and use of AI for good and for all
European approach to AI Inclusive: composition of AI HLEG, EU AI Alliance Comprehensive: addressing competitiveness, socio-economic challenges & legal and ethical issues Ethical: in line with European values
Coordinated action plan on AI – Maximise investments National AI strategies: By mid-2019 all Member States should have their own strategies in place A new European AI public-private partnership A new AI scale-up fund Developing and connecting world-leading centres for AI
Europe has strong assets World-leading position in robotics, especially for professional users Performant industrial and services sectors & strong business-to-business applications World-leading research centers
Boosting the EU's technological and industrial capacity: 2018 - 2020 € 1. 5 billion EC investments into AI in 2018 -20 Support for basic (explainable AI, data-efficient AI) and industrial research (health, transport, agriculture, manufacturing) AI-on-demand platform Network of AIfocused Digital Innovation Hubs (DIHs) Strengthening AI excellence centres
Planning of AI-related investments in next multiannual financial framework (2021 -2027) Digital Europe Programme: - € 2. 5 billion on AI - Common data spaces, AI-on-demand platform, large-scale testing facilities, reinforcing AI excellence centres, AI-focussed Digital Innovation Hubs Horizon Europe (H 2020 successor) - € 100 billion overall - € 53 billion for "global challenges and industrial competitiveness" including AI as part of the "Digital and Industry" cluster - € 16. 6 billion for European Research Council, € 10. 5 billion for European Innovation Council
Making more data available An updated Directive on public sector information Guidance on sharing private sector data in the economy (including industrial data) An updated Recommendation on access to and preservation of scientific information A Communication on the digital transformation of health and care, including sharing of genomic and other health data sets
Coordinated action plan on AI – Create European data spaces Identify public data sets Develop a database with cancer images Support the development of industrial data platforms
Social acceptance of robotics and AI 2012 2014 2017 General favourability 74% 61% Steal jobs 70% 72% 36% 44% 48% 35% Could do my job Being comfortable with a robot at work 48%
Do we want full transparency of algorithms? Or should we trust experts that have certified a product?
Coordinated action plan on AI - Skills Exchange best practices on upskilling Include the skills dimension in national AI strategies Make full use of the Blue Card system Develop a Ph. D component as part of the efforts to strengthen European AI excellence centres
Preparing for socio-economic changes Set up dedicated (re-)training schemes in connection with the Blueprint on sectoral cooperation on skills Gather input from experts to anticipate changes on the labour market Support Digital Opportunity Traineeships (2018 -20) in advanced digital skills for students and fresh graduates Ask social partners to include AI in their joint work programmes Suggest broadening the scope of the current European Globalisation Adjustment Fund to digitisation/automation
The AI High-Level Expert Group 52 Experts 40% women Working Group 1 Chair: Pekka Ala-Pietilä 23 companies Product: draft AI Ethics Guidelines Vice-Chair: Nozha Boujemaa Working Group 2 Product: Policy & Investment Recommendations 19 academia Vice-Chair: Barry O'Sullivan 10 civil society
The European AI Alliance Joint reflection on the future of AI in Europe Full mobilisation of all stakeholders needed: industry, academia, civil society Supported by high-level expert group on AI and an online platform Goal: Making it the world-wide reference platform for thinking and reflection on AI
Draft AI Ethics Guidelines • Consultation to gather feedback: • DDL 01/02 • Everyone can participate via EU AI Alliance: https: //ec. europa. eu/futurium/en/ai-alliance-consultation • Next steps: • • Finalisation in March & presentation in April And then? Not yet established but stakeholders will in principle be able to endorse the guidelines formally (self-regulation) “Trustworthy AI brand? ” Periodic evaluation & adaptation
What are the ethical challenges of artificial intelligence? Transparency Bias Job losses & inequality Security Humans staying in control
Draft AI Ethics Guidelines • Chapter 1: • • Chapter 2: • • • Fundamental Rights & 5 Core Principles • Beneficence (do good), Non-Maleficence (do no harm), Autonomy of humans, Justice, and Explicability. 10 Requirements for Trustworthy AI • Accountability, Data Governance, Design for all, Governance of AI Autonomy (Human oversight), Non-Discrimination, Respect for Human Autonomy, Respect for Privacy, Robustness, Safety, Transparency. Technical & Non-technical methods Chapter 3: • • Non-exhaustive assessment list based on questions Use cases: healthcare diagnosis, self-driving car, insurance premiums, law enforcement
Coordinated action plan on AI – Develop trustworthy AI Draft AI ethics guidelines by April 2019 Reach out to international partners and promote the AI ethics guidelines Organise an international ministerial meeting on AI in April Publish a report on the broader implications of AI for the liability and safety frameworks
- Slides: 19