Warning over brain drain of European AI talent to industry

European countries need to urgently improve conditions for researchers working in artificial intelligence to stem a brain drain of talent moving from academia to industry, sector leaders have said.

The warning was issued after a major new report on the AI research landscape suggested that US technology companies were enticing top researchers away from universities at an increasing rate.

The Elsevier report – Artificial Intelligence: How Knowledge is Created, Transferred and Used – shows that, globally, research output on AI grew by more than 12 per cent annually from 2013 to 2017 compared with just 0.8 per cent a year for all subjects. China was responsible for almost a quarter of publications in that period, against 17 per cent for the US, the country with the next biggest output.

Taken as a whole, Europe is still ahead of China on output (although the report says that it is bound to be overtaken in the “near future”) but it is the findings on academic mobility that could be the most worrying for policymakers on the continent.

While China is operating in “relative isolation” on AI, with a large share of its researchers staying in the country, there has been a trend of more scholars moving out of Europe.

Over the past 20 years, the share of academics leaving the continent for more than two years was a percentage point higher than inward migration, while in the US and China there was a small net inward movement of talent.

More detailed figures show that the biggest net flow of researchers between regions and sectors over the past five years has been towards US industry from international academia. Europe, meanwhile, had a net loss of academic talent to companies both abroad and within the continent.

Maria de Kleijn-Lloyd, senior vice-president of analytics services at Elsevier, said that the findings should be “a call to action for Europe” on working conditions in industry and academia.

She pointed out that while there are a lot of researchers moving between academia and industry in the US, there is a “really remarkable” culture of joint research working between the sectors, meaning it is easier to switch from one to the other.

This is shown by academic-corporate collaboration representing 9 per cent of US AI research output, much higher than Europe (3.6 per cent) or China (2.3 per cent). Microsoft and IBM also made the top five for the most individual contributions to AI research, with Google further down the list but “ramping up very, very rapidly”.

Ms de Kleijn-Lloyd said that a “major question” for European research is about access to large datasets. A major attraction of the US is the large datasets held by corporations, while China has access to data from its own social media platforms.

Holger Hoos, professor of machine learning at Leiden University, said that he was “deeply concerned” by the brain drain of researchers from European universities, which he said was “progressing at an alarming rate”.

He has co-launched an initiative called CLAIRE, which already has the support of around 1,400 AI experts. It aims to strengthen pan-European cooperation on AI and hopes to establish a central state-of-the-art AI hub similar in scale to the Cern physics lab in Switzerland.

“We need a massive pan-European effort to coordinate and support excellence in AI research and innovation – an effort that generously supports the very best talent in the field and that matches expertise between academia and industry,” Professor Hoos said.

Nick Jennings, Imperial College London’s vice-provost for research and enterprise and AI chair at the institution, said that if the brain drain issue was not tackled, “it is likely to have a significant impact on our ability to train the next generation of talent in this area”.

“Many experienced people and leaders in the field are leaving academia. While this is perfectly understandable from an individual’s perspective, when you look at it from a systemic perspective it is damaging for students, universities and in the longer term the very companies that are now employing such staff.”

He called for “more flexible contracts that allow staff to be employed by both companies and universities at the same time” and for governments to look at salary top-ups for researchers to retain the best talent in academia.

simon.baker@timeshighereducation.com