Brando Benifei MEP (S&D, Italian) and Michael McNamara MEP (Renew Europe, Irish), co-chairs of the parliamentary working group on artificial intelligence, have welcomed the “compromise” represented by the very recent ‘Code of Practice on General Models of Artificial Intelligence’, even though its signature by the major suppliers is still uncertain (see EUROPE 13678/17).
“This code provides a necessary first layer of protection - offering structured and practical guidance to ensure compliance with the AI Regulation, while supporting the responsible development of powerful AI systems”, they write.
According to the two MEPs, the survival and integrity of the Code now depend on its ability to be implemented: “The credibility of the European AI framework depends on the ability of regulators to provide robust oversight, supported by sufficient resources and the political will to audit, investigate and sanction, if necessary.”
“We call on the Commission, the Member States, the AI Office and suppliers to strongly support this framework, to reject efforts to weaken it and to maintain the spirit of cooperation that has made it possible”, they say.
For some players in the sector, the Code still imposes a disproportionate burden on AI providers, and comes too late (see EUROPE 13679/7).
Many are calling for a change in the timetable for implementation, to give the companies concerned “time”. (Original version in French by Isalia Stieffatre)