A crucial meeting is happening today in the European Union (EU) regarding the future of Artificial Intelligence (AI). Key EU bodies – the European Parliament, the Council, and the European Commission – are meeting to discuss the Artificial Intelligence Act (AI Act), a major law that will shape how AI is used across Europe.
Michelangelo Baracchi Bonvicini, the President of Atomium – European Institute for Science, Media and Democracy, had this to say:
If an agreement is not reached this week, this will represent a big missed opportunity for Europe and one of the biggest scandals involving the influence of interest groups in the internal affairs of the EU.
France, Germany, Italy’s Stance on AI: A Divergence from EU’s Vision?
This meeting comes after an agreement between France, Germany, and Italy on a new approach to regulating AI.
The document, seen by Reuters, shows that these countries favor a self-regulated approach, especially for basic AI models, using a set of guidelines. The proposal acknowledges that AI technology itself isn’t dangerous, but it points out the importance of having clear rules for how AI is used.
At the heart of their proposal is the introduction of “model cards” for AI. These cards would provide important details about what an AI model can do and its limitations. Initially, there will be no fines for not following these guidelines, but this step is seen as the beginning of building a stronger set of rules for managing AI technology.
These countries want to go with a more flexible, self-regulated way of managing AI, especially for basic types of AI. However, this approach is quite different from what the EU’s AI Act suggests. The EU Act wants a more uniform and strict set of rules for all EU countries. Although the use of model cards is good for transparency, it might not be as thorough as the EU’s AI Act, which plans for stricter control and rules.
Also, by focusing mostly on simpler AI models, the approach of these countries doesn’t cover as many types of AI as the EU Act does, which includes rules for more complex and risky AI systems.
If these three countries keep pushing their way of doing things, it could lead to talks being derailed today and the agreement being delayed.
Charting the Course: The Evolution and Impact of the EU’s AI Act
The AI Act, first proposed in April 2021 by the European Commission, is a groundbreaking step to ensure that AI technology is used safely and respects people’s rights. It’s about finding a balance between embracing new tech and keeping everyone secure.
Last year, on December 6, the European Council made some big changes to the proposed law. They fine-tuned the definition of AI and made sure private companies couldn’t use AI to unfairly rate people’s social behavior, a practice known as social scoring. They also made the rules simpler to follow.
The European Parliament has also been a key player in shaping the law. In June 2023, they proposed major changes, like adjusting the AI definition to match global standards and banning certain uses of AI for identifying people through biometrics, which is like using technology to recognize faces or other personal features.
Today’s meeting is about working through these complex issues. The main focus will be on agreeing on the types of risks AI poses, how to handle AI that can be used for many things, and ensuring AI development doesn’t harm people’s rights.
This meeting is a big deal because the decisions made here could influence how AI is regulated not just in Europe, but around the world. The AI Act is a key part of the EU’s digital strategy, showing how Europe wants to lead the way in making sure AI is used responsibly.