The Strategist

EC President names 4 conditions to ensure safe AI

11/03/2023 - 05:33

According to Ursula von der Leyen, president of the European Commission, there should be four requirements for safety measures to "understand and mitigate the risks" connected to the application of artificial intelligence (AI).

Photography by Bob Mical
Photography by Bob Mical
She listed them speaking at the first summit on the safe use of AI in the UK.

"First and foremost, we require a robust, autonomous scientific community with the resources necessary to assess AI systems. They require access to the greatest supercomputers and government financing," von der Leyen declared.

Second, "internationally recognized procedures and standards for AI safety testing need to be developed."

Thirdly, "there should be a standardised procedure whereby every significant incident caused by errors or misuse of AI will be reported and acted upon."

The EC chief noted that a "robust international alert system" ought to be put in place as a fourth requirement.

A system of governance that works should be built around these four pillars. A culture of accountability ought to be at the center of everything. The EC President clarified that for private players, this entails a single fundamental principle: the more the AI model's capabilities and the accompanying risks, the greater the duty.