Ahead of the next round of trilogue negotiations on the EU’s Artificial Intelligence (AI) Act, the Computer & Communications Industry Associations (CCIA Europe) sent a letter to the Spanish Presidency and the European Parliament’s negotiating team.
With only a short time left to reach a deal, CCIA Europe believes the EU co-legislators should avoid undermining the Act’s original risk-based approach by suddenly introducing an expansive asymmetric approach to regulate foundation models and general-purpose AI (GPAI).
As the letter makes clear, the AI Act should not include any asymmetric obligations that inappropriately target a few providers or models, irrespective of risk or use.
“It is still very early days for AI technology, and rapid developments are yet to come. The market is already fundamentally different than one year ago with many new players and great leaps in what AI can do. Any asymmetric regulation is likely to become outdated within a few years, if not months.”
“There are clear examples that smaller AI models, and models from smaller companies, can have the same or higher impact as larger models and companies. It all depends on how they are used, not necessarily by whom. There is simply no evidence of any correlation between these risks and the size of providers.”
