If Europe is to make the most of AI, it must find a way to bring together stakeholders from across the continent. One potential source of leadership is the northern European group of nine “digital front-runners,” namely Belgium, Denmark, Estonia, Finland, Ireland, Luxembourg, the Netherlands, Norway, and Sweden. These “DF9” are ahead on AI infrastructure and policy environment, and have the potential, given the right conditions, to become Europe’s AI locus, driving adoption and helping Europe maximize the technology’s impact.
Since the start of the last decade, momentum across AI research, education, employment, and investment has increased significantly. The number of published papers doubled from 2010 to 2018, while AI patent applications tripled. Enrollment in AI and related studies has expanded rapidly, both at the university level and online, and AI has become the most popular specialization among computer science PhDs. The majority of AI PhD graduates now get hired to industry rather than academia, with leading companies offering researchers up to ten times their academic salaries, amid intense competition for talent. Start-ups are at the forefront of AI innovation—some of the most valuable companies today did not exist just a few years ago. Global investment in AI start-ups grew from €1.2 billion in 2010 to €36.7 billion in 2018.
Many governments are developing or rolling out national strategies that both reflect their AI ambitions and aim to tackle perceived challenges, for example, in relation to accountability for decisions made by algorithms, data protection, and potential shifts in labor markets. The European Commission has published a Europe-wide plan aimed at ensuring competitiveness and encouraging knowledge sharing. It has also asked individual governments to develop their own strategies.
Early indications suggest that the momentum might have been amplified during the COVID-19 outbreak. The benefits of the technology have become more tangible as governments and companies have used AI tools to combat the virus. For instance, two tech companies in the Netherlands have trained an AI algorithm to detect COVID-19 via X-ray scans, helping more than a hundred hospitals ramp up testing. 1 1. CAD4COVID, Delft Imaging, 2020, delft.care. As with the SARS outbreak in 2003, it may be that the COVID-19 pandemic could be a catalyst for accelerating technology (and in this case AI) adoption, both in combatting the disease and supporting business activities.
In formulating AI strategies, China and the United States have prioritized three pillars: early-mover advantage, private and public investment, and development of large technology sectors and talent pools. In 2017, China published a document that highlighted its ambition to establish global leadership by 2030. 2 2. “Notice of the State Council issuing the new generation of artificial intelligence development plan,” The State Council of China, State Council Document No. 35, July 2017. The United States first issued a national AI report in 2016 and followed up with an executive order in 2019 that emphasized maintaining leadership. 3 3. The national artificial intelligence research and development strategic plan, National Science and Technology Council, Executive Office of the President of the United States, October 2016 and June 2019, whitehouse.gov. 4 4. Donald J. Trump, “Executive order 13859: Maintaining American leadership in artificial intelligence,” Government of the United States, February 11, 2019, whitehouse.gov. Both countries aim to lead across all sectors, rather than focusing on specific strengths.
On a governmental level, China and the United States are investing significantly in R&D. The US government has budgeted around €4.5 billion for 2020, matching its allocation in 2019. While China does not disclose its national budget, the cities of Shanghai and Tianjin plan to invest around €13.5 billion over the next decade. By comparison, the European Union invested around €275 million annually from 2014 to 2017 under its Horizon 2020 program.
China and the United States are home to most of the world’s top tech companies, and private-sector investment in the United States dwarfs that in most other regions. Between 2012 and 2018, the United States invested 20 times more in AI and big data than Europe. 5 5. PitchBook: Venture Capital, Private Equity and M&A Database, 2020, pitchbook.com. In 2018, it attracted 46 percent of total private investment in AI, while China had 36 percent and Europe had 8 percent.
During the COVID-19 crisis, China and the United States have taken a lead in using AI for automated diagnosis, exploring vaccines and treatments, and predicting disease development in society and individuals. For example, in the United States, a machine learning algorithm was developed to predict when infections will slow down in each country and quantify the impact of quarantine measures. 6 6. George Barbastathis and Raj Dandekar, Quantifying the effect of quarantine control in COVID-19 infectious spread using machine learning, medRxiv—the preprint server for health sciences, April 2020, medrxiv.org.
During the COVID-19 crisis, China and the United States have taken a lead in using AI for automated diagnosis, exploring vaccines and treatments, and predicting disease development in society and individuals.
Tech clusters are responsible for an outsized proportion of AI innovation, entrepreneurial activity, and economic growth, and the effect is self-reinforcing: better research through more talent and capital leads to higher levels of commercialization, which fosters investment and, in turn, attracts new talent. San Francisco’s greater Bay Area cluster was the recipient of 46 percent of all AI venture-capital (VC) funding in the United States in 2018 and was responsible for 52 percent of VC-backed exits. 7 7. The global startup ecosystem report 2019, Startup Genome, May 2019, startupgenome.com. 8 8. Justin Fox, “Venture capital keeps flowing to the same places,” Bloomberg, January 8, 2019, bloomberg.com. 9 9. “18 charts to illustrate US VC in 2018,” PitchBook, January 28, 2018, pitchbook.com.
McKinsey research shows that the United States is home to 18 of the top 25 clusters, while only three European and four Asian cities rank in the top 25. Outside of London, none of the top 25 clusters are in Europe.
While China and the United States are moving forward with determination, Europe has the resources at its disposal to keep pace. The continent has a large number of strong incumbent industries, the world’s largest single-market area, a sturdy legal framework, excellent public services, and many companies and small and medium-size enterprises that are leaders in their fields. Europe also boasts high-quality education and research capabilities; it has more professional developers than the United States and has been the most prolific publisher of AI papers over the past 20 years. 10 10. Raymond Perrault et al., The AI index 2019 annual report, AI Index Steering Committee, Human-Centered AI Institute, Stanford University, December 2019, hai.stanford.edu.
On a policy level, the European Union is leading coordination among its member states for investment and data sharing and is also developing an ethical and legal framework. The European Union’s “Coordinated Plan on Artificial Intelligence” encourages the use of AI to solve some of the world’s most pressing challenges, including disease, climate change, and cross-border crime. 11 11. “COM (2018) 795: Coordinated plan on artificial intelligence,” European Commission, December 7, 2018, ec.europa.eu.
Across Europe’s diverse economies, some countries have ramped up investment in AI, while others are a little behind. The United Kingdom, Germany, and France are home to Europe’s largest tech hubs, supported by dedicated policies and fiscal initiatives such as R&D tax credits. With the help of an effective AI strategy, the United Kingdom has become the world’s fourth largest market for investment, after the United States, China, and Israel, and is home to some of the biggest names in the business. 12 12. PitchBook: Venture Capital, Private Equity and M&A Database. Germany, meanwhile, is seeing a wave of innovation, with industrial giants leading the world in patent applications in some categories. France has instigated a national AI strategy, with ambitions that include wider access to big data, deeper research, and the development of plans to cater for the impact of AI on labor.
A proportion of European AI activity is associated with the DF9. These countries boast high levels of tertiary education, advanced digital infrastructures, populations that are amenable to new technologies, and strong corporate buy-in—some 62 percent of large companies in the nine countries have adopted AI in some aspect of their daily operations. The DF9’s history of collaboration, based on their strong cultural and business ties, is a good fit for the concerted efforts required to develop AI applications. Together they have the potential to become Europe’s AI locus, driving adoption, encouraging research, and maximizing the technology’s impact.
The DF9 are all relatively advanced when it comes to AI readiness, a strong predictor of the ability to generate employment and growth. The best positioned of the group is Finland, which has embraced digital through multiple private and public initiatives and was the first European country to implement a national AI strategy. Finland performs well across all key indicators for AI readiness, while Denmark, Luxembourg, the Netherlands, Norway, and Sweden all show strong potential to take AI forward. The DF9’s key strengths in terms of AI readiness include high levels of technology adoption potential, a solid digital and data infrastructure, and a supportive policy environment.
The DF9 are all relatively advanced when it comes to AI readiness, a strong predictor of the ability to generate employment and growth.
If the DF9 adopt AI at scale, the potential economic impact could be as high as €42 billion annually (or 1.4 percent of GDP), McKinsey estimates show. Labor substitution and innovation are likely to be the primary drivers, accounting for around 70 percent of the uplift. The high level of labor substitution stems from AI’s potential as an enabler of automation. Tasks that may be automated represented 44 percent of working hours in 2017, based on McKinsey estimates. Meanwhile, AI may fuel demand for some products and services. As open and digitally advanced economies, the DF9 are expected to benefit from increased data flows and reinvestment of income generated by AI-enabled wealth creation.
There also may be some less favorable impact. As the DF9 economies transition to new ways of working, companies and employees will need to adapt. Firms will be required to hire the right employees, ensure proper data management, and introduce new technological infrastructures. Employees will likely need to re-skill, which could entail temporary unemployment. Still, in the longer run, we expect AI will have a neutral impact on jobs, with as many created as are rendered obsolete.
Current AI technologies have the potential to unlock value across all sectors and all parts of the value chain in the DF9 economies. 13 13. Based on an analysis of 400 use cases across industries; for each use case, the annual value potential of applying AI is estimated. However, retail and travel have the most potential relative to their size. Other potential applications include price optimization in finance, yield enhancement in manufacturing, next product to buy in retail, and predictive maintenance in energy. AI techniques are likely to generate value above and beyond traditional analytics, with the uplift ranging from 30 percent to 128 percent, depending on the industry.
AI has the potential to deliver social and welfare benefits in the DF9 and globally, with possibilities ranging from increased environmental sustainability to more meaningful jobs focused on creativity and social skills. It is already starting to have an impact, with multiple use cases across the United Nation’s 17 Sustainable Development Goals. Likewise, the COVID-19 outbreak has demonstrated AI’s potential in delivering healthcare and safety benefits, with DF9 countries using it to support diagnosis, control, and forecasting.
Within the DF9, the most fertile areas for AI development are probably healthcare and the work environment. To get a sense of the potential impact, a useful metric is “GDP equivalent.” Preliminary results on this basis quantify the annual GDP-equivalent impact as 0.4 percent of GDP. Potential benefits will include better health, fewer working hours, and increased leisure time, while potential drawbacks may include employee stress, inequality, and unemployment.
Europe is currently at risk of falling behind in the race on AI, which may have an effect on its relative productivity performance in the years ahead. However, the DF9, which are European leaders on key readiness indicators, may provide a boost to Europe’s standing. To deliver, they need to overcome a number of challenges, including a lack of skills (both technical and managerial), regulatory uncertainty (for example, relating to data privacy and liability), confusion over the business case (often as a result of the absence of a clear strategy), competing investment priorities, and, in some cases, a shortfall of high-quality data and supporting infrastructure. In addition, a significant level of private-sector investment will be required.
To address challenges and meet the needs of a fast-growing technology, we believe the group could focus on the following five priorities:
AI is set to become a game changer in the public and private spheres. The technology has the potential to boost productivity and help humans work better, faster, and with more perspective in almost every activity. The DF9 are therefore presented with a historic opportunity. The nine countries together have the technical know-how, investment resources, and digital culture to take a lead on AI, and to put in place the frameworks, structures, and research centers that will drive adoption for both corporate profit and social good. It is incumbent now on the DF9 to move forward and, in the process, ensure that Europe reaps the benefits of AI in the years ahead.
This content was originally published here.