The statement addresses concerns about the direction of the latest proposals to regulate General Purpose AI (GPAI) systems and AI Foundation models. The current proposals do not take into account the complexity of the AI value chain and are not consistent with the AI Act’s technology neutral risk-based approach.
- The proposed classification of GPAIs and AI Foundation Models as “highly capable” or as having “high impact” is based on criteria that are not clearly linked to the level of risk that an AI system developed with GPAIs and Foundation Models may pose.
- The AI Act should formalize a mechanism to ensure the sharing of relevant and necessary information between Foundation Model providers and deployers, for instance about model capabilities and limitations.
- The use of copyrighted data to train AI systems is already covered by the existing EU legal framework and introducing any additional requirements through the AI Act is not a viable legal solution.
The following quote can be attributed to Karina Stan, Director of EU Policy of the Developers Alliance:
We strongly call on the EU lawmakers to stick with the high-risk approach for the AI Act, and maintain it as a product safety regulation. In this sense, any provisions regarding General Purpose AI and Foundation Models should consider the variety of applications and of value-chains. Instead of regulating the technology itself, which is rapidly evolving, the AI Act should provide a strong risk management framework. This should include support for developers and deployers downstream to receive, from the Foundation Models providers, all the necessary information that will allow them to be compliant with the AI Act.
About The Developers Alliance
The Developers Alliance is the world’s leading advocate for software developers and the companies invested in their success. Alliance members include industry leaders in consumer, enterprise, industrial, and emerging software development, and a global network of more than 75,000 developers.
Head of Marketing and Communications