Proposed AI and biometrics restrictions by EU draw mixed, muted reaction
Reaction is rolling in to the European Union’s attempt to coral artificial intelligence with regulation, which includes strict new proposed rules for the use of biometrics in public spaces by law enforcement.
The rules aim for a middle path between China and the U.S., in terms of government involvement, the Financial Times writes. Columbia University EU Law Professor Anu Bradford describes the new rules as an attempt to re-establish competitiveness in AI without compromising on human rights
The use of biometrics and predictive algorithms at borders is still allowed, if the right regulatory hoops for ‘high-risk’ AI are jumped through, the Times reports.
Representatives from European Digital Rights criticized the rules for not providing enough recourse mechanisms for individuals adversely affected by AI.
The role of trust is placed at the center of the EU’s bet, however, Bradford says, as AI systems seen as trustworthy could have an advantage in the broader market.
“It’s going to make it prohibitively expensive or even technologically infeasible to build AI in Europe,” Center for Data Innovation Senior Policy Analyst Benjamin Mueller told the Wall Street Journal. “The U.S. and China are going to look on with amusement as the EU kneecaps its own startups.”
Some lobbyists told the Journal they were relieved at the proposal, however.
Diginimica points out that the European Commission rules fit most AI applications into the ‘minimal risk’ category, just below chatbots and other ‘limited risk’ uses. Remote biometric systems get even stricter treatment than other ‘high-risk’ systems, though they are not doomed to the ‘unacceptable risk’ category.
The publication rightly points out that competing estimates of compliance costs will soon be presented as arguments for and against the EC’s proposal. Further, the risk categories may be too vague for regulators with little understanding of the technology to work with.
European regulations like GDPR have influenced regulations passed since in other jurisdictions, however, as a lawyer interviewed by Wired notes. The Observer, however, points out that only one of 13 bills and resolutions on AI made it through state legislatures last year.