European industry has no-one to blame but itself for the slow pace of detailed AI standards that are core to implementing the EU’s AI Act, a leader of a standards drafting process said today.
The EU’s AI Act relies upon detailed technical standards to turn broad principles into concrete rules and procedures for AI developers to follow.
But standards development has been slow, helping to fuel repeated calls from industry and EU governments to delay the law’s application.
Piercosma Bisconti, a project leader drafting an “AI trustworthiness framework” – to technically specify the Act’s requirements for logging, transparency, human oversight, accuracy and robustness – has publicly called out companies for not engaging with the process.
“These are precisely the actors who should be contributing, and they are not,” he wrote on LinkedIn on Monday – referencing European tech industry players which his post accuses of being responsible for a delay that the same companies have used to justify calls for freezing the law’s application.
He added that “EU industry is barely at the table”.
Bisconti, the co-founder and scientific director of an Italian AI company (Dexai), is doing work for the European Committee for Standardization (CEN) and European Committee for Electrotechnical Standardization (CENELEC) on standards for AI transparency and human oversight, after also working on the EU’s code of practice for general purpose AI models.
Bisconti is specifically critical of companies that have been involved in the so-called “AI champions initiative” – which has called for “stopping the clock” on the AI Act.
The more than 110 signatories to the initiative include high profile companies including Airbus, Siemens, Spotify and SAP, along with smaller players and European startups.
“If companies are concerned about delays, they should reflect on their own role,” he wrote. “Harmonized standards do not write themselves.”
(nl)
