noun_Email_707352 noun_917542_cc Map point Play Untitled Retweet Group 3 Fill 1

8 Enablers For Europe's Trustworthy Artificial Intelligence

AI Experts and practitioners from all across Europe gathered on 26th June 2019 at the first European AI Alliance held in Brussels, Belgium. Here are some of my reflections and key assessments.

Emmanuel Raj / July 05, 2019

AI Experts and practitioners from all across Europe gathered on 26th June 2019 at the first European AI Alliance held in Brussels, Belgium. I had the privilege to attend the first European AI Alliance. Here are some of my reflections and key assessments.

Many questions were raised and answered in order to consolidate a framework and envision a promising future for European AI. An intriguing question was raised during the event "How is European AI different? How will we catch up with America and China?"

When it comes to Artificial intelligence, America is known for its winning attitude, China for finding the middle ground for innovation and Europe for its values. We should humble ourselves and listen more.

Loubna Bouarfa, Member of AI HLEG

With this hunger and zeal to innovate and learn, Europe's AI leaders consolidated and agreed upon a vision to build Ethical and Trustworthy AI systems. Since last year AI HLEG, together with AI Experts has consolidated a foundational framework(Ethics Guidelines) to kickstart united European AI efforts. And recently at the first European AI Alliance Policy and investment recommendations for Trustworthy AI were consolidated in a hope to foster AI innovation, growth and sustainability for Europe and its citizens. Here is my assessment on our current realities, solutions and action points. These are the 8 enablers for Europe's Trustworthy Artificial Intelligence suggested by the European AI Alliance.

Image Source: Policy and investment recommendations for Trustworthy AI(Pg: 25)

1. Data & Infrastructure

Current state: Data is an indispensable raw material for developing AI. With data volumes growing at 61% per annum in Europe, four times more data will be created until 2025 than that which exists today. Ensuring that individuals and societies, industry, the public sector as well as research and academia in Europe can benefit from this strategic resource is critical, as the overwhelming majority of recent advances in AI stem from deep learning on big data. A safe, secure and high-quality data infrastructure would enable Europe to better develop and train AI systems, which in turn can be steered towards applications that can facilitate the Sustainable Development Goals.

Solution and action(s):

  • European AI Alliance will focus and drive setup of national and European data platforms for AI that include all necessary tools for data governance, annotation, and storage, next-generation networks, analytics software and, most importantly, datasets through a structural and investment fund.
  • With security and privacy as a focus, a fundamental rights-based personal data infrastructure as put forward in the GDPR will be fostered and its enforcement will be ensured.

2. Skills & Education

Current state: Thanks to the European education system Europe has been in the forefront of producing high-quality AI talent but the quantity is not enough to keep up for the change. A lack of AI-related skills (including technical AI skills, skills for dealing with AI and managerial capabilities to extend AI in business) has been identified as the most important barrier to AI adoption in Europe. AI talents with expert knowledge are required who are capable of driving, managing and conducting AI activities in their institutions and organisations. In addition to the lack of Skill AI talent, brain drain of existing skilled workers has been eminent (Especially to US and China).

Solution and action(s):

  • Redesign education system from pre-school to higher education - Investments will be made to set up recommendations and incentives to adopt national education systems to strengthen children in human-centric key skills as (i) cognitive competencies like problem-solving, process and quality monitoring, critical thinking, judgement and creativity, etc. (ii) socio-cultural competencies like empathising, leading, persuading, envisioning, etc. (iii) entrepreneurial and innovation competencies.
  • Up-skill and Re-skill the current workforce - (i)Develop and implement a European Curriculum in AI in collaboration with the European Excellence Centre of Trustworthy AI, (ii)Increase disciplinary mobility between AI-associated curricula and (iii)Mainstream and include skills related to data and AI in all academic disciplines and professional fields to increase the potential of areas where AI applications can be developed.
  • Create stakeholder awareness and decision support for skilling policies.
  • Encourage more female talents into the field of AI and related subjects.

3. Governance & Regulation

Current state: Ensuring an appropriate framework that maximises the benefits of AI and that prevents and minimises its risks is no easy task. To deliver trustworthy AI for Europe, independent and meaningful oversight mechanisms need to be established, and an expansion of the institutional capacities, expertise and instruments of policymakers is needed. Yet little evidence is available to inform policy-making, due to the novelty of the technology, the lack of thorough and systematic understanding of its impacts and associated business models, and the unpredictability of its uptake, development and evolution even in the short term. Bearing this in mind, here below are some proposed focuses and solutions that are already considered by policy-makers at EU and national level.

Solution and action(s):

  • Ensure appropriate policy-making based on a risk-based and multi-stakeholder approach.
  • Evaluate and potentially revise EU laws, starting with the most relevant legal domains.
  • Consider the need for new regulation to ensure adequate protection from adverse impacts.
  • Consider whether existing institutional structures, competencies and capacities need revision to ensure proportionate and effective protection.
  • Establish governance mechanisms for a single market for trustworthy AI in Europe.

4. Funding and Investment

Current state: The forecast of the AI worldwide market value shows a fast growth, with AI reaching $118 billion by 2025 from $9.5 billion in 2018. Europe currently attracts only ~11% of global VC funding, with ~50% going to the US and the rest mostly to China. An assessment of the economic activity growth due to AI until 2030 shows that the value at stake for Europe is significant: if no actions are taken, the EU28 will suffer a deterioration of its innovation capital, which would result in a loss of €400 billion in cumulative added value to GDP by 2030.

Solution and action(s):

  • Enable an open and lucrative climate of investment that rewards trustworthy AI.
  • Multi-Stakeholder Alliances that will enable AI ecosystems on a sectoral basis.
  • Dedicated, significant and long-term research funding for fundamental and purpose-driven research on AI to maintain the competitiveness of European companies and address relevant societal challenges.

5. Private Sector

Current state: Despite the sizeable impact that European businesses expect from AI, only a small fraction of them use AI actively in industrial and commercial operations. In 2018, three-quarters of European businesses did not adopt AI in any shape or form, just under a single quarter were in piloting and testing phases and reported difficulties of scaling, and only 2-3% of companies incorporated AI across their whole organisation.

Solution and action(s):

  • Boost the uptake of AI technology and services across sectors in Europe. Earmark significant resources in the InvestEU programme to support the transformation of European enterprises towards AI-enabled solutions.
  • Foster and scale AI solutions by enabling innovation and promoting technology transfer.
  • Set up public-private partnerships to foster sectoral AI ecosystems.

6. Civil Society

Current state: AI presents a promising means to enhance individual and societal well-being and the common good, as well as driving progress and responsible innovation. Yet it also carries risks for humans and societies, which need to be identified and addressed. It is therefore essential that individuals gain awareness, knowledge and understanding of the capabilities, challenges and limitations of AI systems. This requires research, new monitoring and measuring mechanisms, as well as measures that can protect individuals and society from potential adverse impact generated by the technology.

Solution and action(s):

  • Empower humans by increasing knowledge and awareness of AI. Encourage Member States to increase digital literacy through courses (e.g. MOOCs) across Europe providing elementary AI training. Eg: Elements of AI (Finland).
  • Protect the integrity of humans, society and the environment - Eg: Refrain from disproportionate and mass surveillance of individuals, Introduce a mandatory self-identification of AI systems.
  • Promote a human-centric approach to AI at work.
  • Leave no one behind (Eg: Children).
  • Measure and monitor the societal impact of AI.

7. Public Sector

Current state: The EU's commitment towards the modernisation of public administrations was already confirmed by the Ministerial Declaration on e-Government adopted in Tallinn on 6th October 2017. Deploying AI systems can help governments make better evidence-based policy-making decisions, deliver better services to individuals, groups and organisations by reducing internal costs, increasing programme effectiveness, and enhance quality. This should not lead to a lower quality of human relationships within public services or a reduction of such services.

Solution and action(s):

  • Provide human-centric AI-based services for individuals.
  • Approach the government as a platform, catalysing AI development in Europe - (i) Foster digitalisation by transforming public data into a digital format. (ii) Provide data literacy education to government agencies. (iii) Create European large annotated public non-personal databases for high-quality AI.
  • Make strategic use of public procurement to fund innovation and ensure trustworthy AI.
  • Safeguard fundamental rights in AI-based public services and protect societal infrastructures.

8. Research and Academia

Current state: The popular narrative around artificial intelligence research is that it's mainly a war between China and the United States but Europe has been releasing more research papers than either the US or China.

New data released on (Dec. 12) by the AI Index, a project to track the advancement of artificial intelligence, shows a trend of Europe releasing more papers than either the US or China. If the current trend continues, China will soon overtake Europe in the number of papers published. The number of papers out of China grew 17% in 2017, compared to a 13% increase in the US, and 8% in Europe. In order to maintain and Ensuring World-Class Research Capabilities below solutions are proposed.

Solution and action(s):

  • Develop and maintain a European strategic research roadmap for AI.
  • Expand AI research capacity in Europe by developing, retaining and acquiring AI researchers.
  • Build a world-class European research capacity.
  • Increase and streamline funding for fundamental and purpose-driven research.

Investments and efforts are foreseen for these enablers for the next 5 years in Europe. With a promising future comes responsibility for diligent effort. Looking forward to being part of this promising endeavour. Thank you for reading. Please let me know your thoughts.


Emmanuel Raj
Machine Learning Engineer

Emmanuel is an AI expert with 4+ years industry experience. Machine Learning Engineer at Tieto and Member of European AI Alliance at European Commission, he is passionate about democratising AI, bringing research & academia to industry and engineering end-to-end Machine Learning systems that improve human efficiency and quality of life.


Emmanuel Raj

Machine Learning Engineer

Share on Facebook Tweet Share on LinkedIn