American tech companies will soon need to meet new requirements in the European Union regarding artificial intelligence and sharing data with smaller rivals, as the bloc seeks to assert its “technological sovereignty” from the U.S. and China.
EU regulators unveiled plans Wednesday aimed at placing more restrictions on machine learning-enabled technologies in fields ranging from public surveillance cameras to cancer scans and self-driving cars.
The legislation, to be drafted by the end of 2020, is also likely to home in on what the EU has learned from antitrust cases against Alphabet Inc.’s Google and ongoing probes into Amazon.com Inc. and Facebook Inc.: how these platforms allegedly use data to quash smaller rivals. These technology giants—which some U.S. politicians such as Democratic presidential candidate Elizabeth Warren want to regulate as public utilities—are in the firing line of the coming EU legislation.
One remedy under consideration is to oblige platforms to share data with smaller rivals, especially when it comes to consumer behavior regarding the products sold by those competitors.
Spokespeople for Amazon, Google and Facebook declined to comment on the proposals.
EDiMA, an industry group representing big platforms including Google, Amazon and Facebook, welcomed the proposals as a “starting point” and said it would engage with the European Commission, the EU’s executive body, to address some of the shared concerns.
Liability rules regarding the content shared on online platforms are also due to be spelled out in the Digital Services Act by the end of the year. EU Commissioner Thierry Breton said Monday that a pledge by Facebook’s CEO Mark Zuckerberg to accept a degree of liability for content posted was insufficient, because he didn’t address the issue of dominant platforms having responsibilities toward smaller players on the platform.
The rules will also come with new restrictions for U.S. and Chinese companies that develop machine-learning-enabled technologies, particularly when they handle sensitive data such as medical records or facial images.
Restrictions are likely to be placed on the use of facial recognition tools for mass surveillance, to limit the number of individuals targeted. Human oversight and disclosure requirements on which data sets are used by AI will be put on all companies operating in the EU.
The European Commission said Europe was behind the U.S. and China in terms of consumer-oriented applications and platforms. However, it hopes to attract public and private investments of €20 billion ($21.6 billion) a year in a bid to keep up in the industrial and public sectors with the use of big data and AI.
“We recognize we missed the first battle, the battle of personal data,” Mr. Breton said. “Europe has everything it takes to lead the ‘big data’ race, and preserve its technological sovereignty.”
If a technology tested in Europe proves to be too opaque or fails to comply with the rules in place, regulators may order the company to reboot the AI and make it learn from scratch based on different data before rollout.
In December, EU Commission Vice-President Margrethe Vestager was promoted to a dual role of enforcing competition law and shaping new rules for the tech sector. After fining Google over $9 billion for anticompetitive behavior over the past three years, Ms. Vestager admitted that fines weren’t working and a broader regulatory approach was needed to change the behavior of big tech.
“We have the fines to punish past behavior, we have the ‘cease and desist’ to stop what you are doing,” she said, adding that the commission was considering several remedies to restore competition.
One measure under consideration is for platforms found guilty of anticompetitive behavior to be ordered to share data with smaller players, so that the marketplace rebounces after anticompetitive behavior, she said.
Another option being considered is ensuring that companies have access to analytics about the online behavior of their customers—data that is held but not always shared by platforms who host their business.
Ms. Vestager said she had been approached by representatives of the financial sector who complained that they have data-sharing obligations that big platforms didn’t. Discussions about establishing reciprocity in data-sharing obligations between platforms and the financial sector were at a very early stage, she said.
“She has learned that data is the fundamental issue and that what she’s done so far has not worked,” said Thomas Vinje, a partner at law firm Clifford Chance. Regulating how platforms use data “is the natural thing to do, but how this will be done still needs to be defined.”