Artificial Intelligence: 18% of the world's top-tier researchers are European, but only 10% work in Europe

The World Economic Forum's Jayant Narayan flagged up the figure in a debate on Artificial Intelligence at the European Economic and Social Committee's July plenary, saying that to compete Europe needs to invest in medium- to long-term actions to create an ecosystem where AI can thrive.

The EESC's last plenary before the summer break hosted a debate on Artificial Intelligence and real values: our digital future with Jayant Narayan, who leads the World Economic Forum's Global AI Alliance. The debate also saw the presentation of three EESC reports focusing on the EU's 2030 digital targets, the economic and social opportunities of digitalisation and digitalisation for all.

EESC president Christa Schweng opened the debate stressing how critical digitalisation was to achieving the vision of "a European Union that prospers economically, is socially inclusive and is environmentally sustainable" on which she has centred her tenure. Data and AI are one of the key components.

"AI continues to change how we live, work, learn and interact", she said. "Now is the time to go beyond a framework and put into practice real values for our digital future", inclusiveness first and foremost.

SMEs: no silver bullets when it comes to AI

Focusing on the challenges facing SMEs looking to deploy AI, Jayant Narayan stressed that while new solutions were starting to emerge such as no-code AI, which allows companies and individuals to work on drag and drop tools to develop their own AI solutions without having fully trained data scientists on their payrolls, this was no silver bullet. An SME which is procuring an AI solution because it does not have the resources to build it in-house needs to consider if it is safe to share its data with the company that is building its AI solution.

Chatbots, predictions for prices and customer outreach were some of the easily deployable solutions that SMEs had started exploring. These were quick wins. But there also needed to be a medium- to long-term horizon focused on empowering SMEs to develop the local, in-house capacity they need to deploy AI. That is where essential aspects such as skills came into play, which, in the case of AI and data science, cannot be developed overnight. Funding and support for innovation played a key role in this process.

Mr Narayan reverted to the importance of the ecosystem during the debate, in response to a question on the reasons why Europe was lagging behind the US and China on AI.

Looking at the US, he said that while there were multiple reasons for this, government support and funding had played a major role, from the emergence of Silicon Valley to the bipartisan approval of the recent Innovation and Competition Act, which entailed a USD 250 billion spend on tech and innovation.

"That funding creates the kind of market where most of your top researchers and data scientists will end up as well. And this is not a short-term solution, but more of a medium- to long-term view of creating an ecosystem where you don't just generate local value, but you also lead in it". He quoted recent studies showing that Europe was home to 18% of the world's top tier AI scientists, only 10% of whom actually work in Europe.

Truly for all?

Mr Narayan also addressed the question of AI and civil society and whether AI development was truly for all or was rather serving the interests of certain groups.

Quoting studies which show that between 30 and 40% of Amazon's recent revenues came though AI-driven recommendations to consumers, Mr Narayan stressed that AI is indeed very pervasive, and that this was why robustness, explainability, trust and transparency were so essential. Progress was being made, on explainability, for instance with AIX 360, to shine some light on machine decision-making. There were also developments in safety, particularly with AI on the edge. But these were cutting edge solutions which not everyone could be expected to have access to.

While there was no lack of principles "out there" – some 200 to 300 have been listed – he said the biggest concern was whether they were truly being put into practice and civil society's interests safeguarded.

Mr Narayan predicted that progress in implementation would come from a mixture of public sector regulation and industry frameworks, including from big tech, which was now recruiting dual training scientists and more diverse teams.

Through dialogue between industry and regulators things would evolve, possibly reaching a point where industry would start taking action even on a voluntary basis.

Digitalisation: the challenges ahead

In the course of the debate Gonçalo Lobo Xavier, rapporteur of the EESC opinion on the EU's 2030 digital targets, said: "we cannot fear AI. We cannot fear progress and digitalisation in our social, economic and private lives. What we should fear is missing the opportunities or not creating the ideal conditions to benefit from technology and progress. The fear that digitalisation and/or AI will steal our jobs is something that is in the mind of Europeans. And that's why regulation is also needed. But we cannot stop time. Yes, it is important to defend jobs and social rights, but the side effects of the slow pace, of waiting to have a legal framework for everything will undermine the future. That's why it is important to be brave".

Antje Gerstein, rapporteur of the EESC opinion on the economic and social opportunities of digitalisation, said: "the COVID-19 crisis has highlighted the digital lag but also the enormous potential for digitalisation in the EU. The benefits of a completed digital single market are forecast to contribute EUR 450 billion a year to the EU's economic output. SMEs are particularly important, they have to attract investment in order to digitalise but many are quite simply exhausted by the pandemic. We also need to understand that digital business models will increasingly shape our world of work. These models need to be designed in a people- and values-oriented way".

Philip von Brockdorff, rapporteur of the EESC opinion on digitalisation for all, said: "digitalisation has to be accessible to all categories of the population. Businesses, workers and people in general need to be prepared for this digital transformation, and for this we need to strengthen our education systems over the coming years. In preparing for digitalisation, it is important to commit sufficient EU funding – tapping effectively into the Just Transition Fund as part of NextGenerationEU, but also the Digital Europe programme. Equally important is that all this is done through a process of social dialogue and with respect for workers' rights, including people working from home".