- No sanctions should be imposed, says joint document
- Commitments would involve large and small suppliers
- The document suggests “mandatory self-regulation through codes of conduct”
BERLIN, Nov 18 (Reuters) – France, Germany and Italy have reached an agreement on how artificial intelligence should be regulated, according to a joint document seen by Reuters, which should speed up negotiations at the level European.
All three governments support “mandatory self-regulation through codes of conduct” for so-called foundational AI models, designed to produce a wide range of outcomes. But they object to “untested standards.”
“Together we emphasize that the AI Act regulates the application of AI and not the technology as such,” the joint document states. “The inherent risks lie in the application of AI systems rather than the technology itself.”
The European Commission, the European Parliament and the Council of the EU are currently negotiating the bloc’s position on this subject.
The paper explains that core model developers should define model cards, which are used to provide information about a machine learning model.
“The model cards will include information relevant to understanding how the model works, its capabilities and limitations and will be based on best practices within the developer community,” the document states.
“An AI governance body could help develop guidelines and could verify the application of the map models,” the joint document says.
Initially, no sanctions should be imposed, the newspaper indicates.
However, if violations of the code of conduct are noted after a certain time, a system of sanctions could be put in place.
Germany’s Economy Ministry, in charge of the issue with the Digital Ministry, said laws and state control should not regulate AI itself, but rather its application.
Digital Minister Volker Wissing told Reuters he was very happy that an agreement had been reached with France and Germany to only limit the use of AI.
“We need to regulate applications and not technology if we want to play in the world’s premier AI league,” Wissing said.
State Secretary for Economic Affairs Franziska Brantner told Reuters it was crucial to exploit opportunities and limit risks.
“We have developed a proposal that can provide a balance between the two goals in a technological and legal area that has not yet been defined,” Brantner said.
As governments around the world seek to reap the economic benefits of AI, Britain in November held its first AI Security Summit.
The German government is organizing a digital summit on Monday and Tuesday in Jena, in the state of Thuringia, which will bring together representatives from politics, business and science.
AI-related issues will also be on the agenda for discussions between the German and Italian governments on Wednesday in Berlin.
Reporting by Andreas Rinke; Written by Maria Martinez; Editing by Mike Harrison, Barbara Lewis and Diane Craft
Our standards: The Thomson Reuters Trust Principles.