Big tech funding AI ethics research to ‘delay and avoid’ regulation

One of Europe’s leading philosophers has called on universities to take the debate about the ethical use of artificial intelligence “out of the hands of the industry” and warned that big technology firms are funding academic research in the area to create the illusion of action and so stall real regulation.

Thomas Metzinger, a professor at the Johannes Gutenberg University of Mainz, said that corporations hoping to profit from AI had created an “industry” of “ethics washing”.

This is an attempt to “organise and cultivate ethical debates in order to delay, postpone, avoid…policymaking and regulation. And that is something you find everywhere right now,” he told university leaders at the European University Association annual conference in Paris.

“There is a high responsibility for European universities to take the ethics debates on artificial intelligence out of the hands of the industry now, and put them on a more neutral and serious academic platform,” he added.

The debate over “ethics washing” has flared up in Germany after the Technical University of Munich (TUM) earlier this year accepted a $7.5 million (£5.7 million) donation from Facebook to found a new AI ethics research centre.

“TUM has done great damage to its reputation in my view, by accepting money from Facebook,” Professor Metzinger said.

“Nobody believes that this is a sincere initiative by Facebook,” he said, adding that there were “all kinds of these strange relationships” between academia and technology companies that compromised research independence.

Christoph Lütge, head of the new centre, said that it had “no obligation” towards Facebook as a result of the gift. “Facebook has really understood that it needs to change its approach,” he said.

“Even if we question the motivations of a company, what counts for ethics is the outcome” and the impact on policy and wider society, he argued.

A spokesman for Facebook said the firm did not wish to comment. 

Professor Metzinger is part of an EU panel of 52 experts that has drawn up ethical guidelines for the use of AI, now being piloted in companies and research institutes.

These guidelines were “strongly industry-dominated” and had “no real normative substance” on “concrete recommendations”, he said. Nonetheless, they were the “best thing we have on the planet right now”, he said, and meant that Europe has taken a global lead on ethical AI.

The EU should spend an eighth of its AI research budget on initiatives exploring the ethical implications of the technology, Professor Metzinger argued.

The conference also heard from Magnus Rattray, director of the University of Manchester’s Data Science Institute, who agreed with Professor Metzinger’s ethics washing concerns. “I think it’s much better if big tech pays tax and then governments fund these kind of centres,” he said.

But he also faced questions over corporate funding of research at the UK’s new Alan Turing Institute, an AI and data science-focused centre with which Professor Rattray is involved.

Industry money was “ringfenced for their particular projects and does not leak into other funded aspects of the programmes”, he said.

Companies possess “a huge amount of data”, making it “quite difficult to envisage an AI and data science national institute that doesn’t work with business”, he added.

david.matthews@timeshighereducation.com