News Hub

France, Germany, Italy reach agreement on AI regulation

Written by Tue 21 Nov 2023

France, Germany, and Italy have formulated an agreement on the regulation of artificial intelligence (AI).

The paper obtained by Reuters is expected to accelerate European negotiations regarding the safe governance of AI.

The three governments are in favour of binding voluntary commitments for both large and small AI providers within the European Union. This self-regulation would be enforced through adherence to predefined codes of conduct.

The three governments are against adopting untested standards of rules. Instead, they said they favour relying on established codes of conduct to regulate AI.

Negotiations are ongoing among the European Commission, European Parliament, and EU Council regarding the bloc’s stance on AI regulation. The European Parliament proposed that the codes of conduct should initially apply only to major AI providers, which are predominantly from the US. 

France, Germany, and Italy expressed concerns around only applying the code of conduct to larger companies. They argued that smaller companies’ use of AI could go undetected or unsanctioned if they are exempt from the codes of conduct. This could develop a lack of trust in the safety and security of these smaller companies, driving business away from these EU providers.

The governments added that the inherent risks lie in the application of AI systems rather than in the technology itself. The paper suggested that developers of foundational models must define model cards, which offer information about a machine learning model. 

“The model cards shall include the relevant information to understand the functioning of the model, its capabilities and its limits and will be based on best practices within the developers community,” said the joint paper.

Alongside model cards, an AI governance body could be established to create guidelines and verify the application of model cards.

Initially, no sanctions will be imposed through non-compliance. However, if violations of the code of conduct are discovered after a specified amount of time, a system to enforce breaches may be created.

Germany’s Economy Ministry and the Ministry of Digital Affairs headed up the topic. The ministries emphasised that AI should be regulated based on its application, rather than through laws and state control.

The EU’s AI Act

Proposed in June, the European Union’s AI Act is expected to set world’s first rules for AI that has the potential to become a global standard. 

If passed, the EU AI Act aims to avoid the safety risks from AI applications and the potential for discrimination without restricting the innovation power of the technology. 

The proposed rules follow a risk-based process depending on the level of risk that AI could create. Any high-risk applications must follow laws and regulations to minimise harm. 

“Together we underline that the AI Act regulates the application of AI and not the technology as such,” said the joint paper.

Whilst the AI Act could be considered a positive step in the direction of AI ethics and governance, multiple organisations like the Future of Life Institute have offered their recommendations for improvement.

Hungry for more tech news?

Sign up for your weekly tech briefings!


Written by Tue 21 Nov 2023

Send us a correction Send us a news tip