European Commission takes cautious approach on AI liability

Daniel Cooper spoke with Global Data Review about the European Commission’s recommended changes to artificial intelligence liability rules. Mr. Cooper says the commission is keen to stress that the conclusions drawn from the report remain those of the expert group only. “One has to assume that the commission wants to be careful to test the waters first and gauge reactions from key stakeholders, like the EU parliament, before openly endorsing the report’s conclusions. This is plainly a public policy debate where the stakes are set very high and the commission wants to tread carefully, as it will impact member state legal frameworks.” He adds that although the report concludes there need be no “reinvention of the wheel” when it comes to the regulation of AI, other remarks about making appropriate adjustments to the law to suit certain high or even lower risk AI applications suggest “that this same wheel will need some serious realigning.”

“The suggestion that existing legal principles will need to be calibrated to AI relying on a ‘risk-based’ approach, whereby higher risk AI applications that could cause greater harm or loss would be subject to a strict liability regime backed by an insurance market while lower risk AI applications receive a different treatment, will undoubtedly be inherently attractive to many – but the devil will be in the detail,” Mr. Cooper said.