The EU has reached an agreement to regulate the development of artificial intelligence

EU co-legislators have reached a “political agreement” on a text that should promote innovation in Europe while limiting the possible excesses of these highly advanced technologies.

“Historic! The EU becomes the first continent to set clear rules for the use of artificial intelligence,” welcomed European Commissioner Thierry Breton at the birth of the project presented in April 2021. From that date, the discussions dragged on. The latest round of negotiations alone, which began on Wednesday afternoon, lasted nearly 35 hours…

This process was influenced at the end of last year by the arrival of ChatGPT, a text generator from the Californian company OpenAI, capable of writing essays, poems or translations in seconds. This system, as well as systems capable of producing sounds or images, revealed the enormous potential of AI to the general public. But also certain risks. For example, the spread of larger-than-life fake photos on social media has highlighted the dangers of opinion manipulation.

This phenomenon of generative artificial intelligence has been included in the current negotiations at the request of MEPs, who insist on special supervision of this type of highly effective technology. In particular, they called for greater transparency of the algorithms and giant databases at the heart of these systems. Member States were concerned that over-regulation would nip their budding champions in the bud, such as Aleph Alpha in Germany and Mistal AI in France, as development costs would be prohibitive.

“Strategic Autonomy”

The political agreement reached on Friday evening must be supplemented by technical work to finalize the text. “We will carefully analyze the compromise we found today and in the coming weeks we will ensure that the text preserves Europe’s capacity to develop its own artificial intelligence technologies and preserves its strategic autonomy,” responded French Digital Minister Jean-Noël Barrot.

The technology sector is critical. “Speed ​​seems to have won out over quality, which could have disastrous consequences for the European economy,” said Daniel Friedlaender, European manager of the CCIA, one of its main lobbies. According to him, “technical work is now necessary” on the essential details. For generative artificial intelligence, the trade-off provides a two-speed approach. All will be subject to rules to ensure the quality of the data used to develop the algorithms and to verify that they do not infringe copyright law. Developers will also need to ensure that the sounds, images and texts created are clearly labeled as artificial.

Hardened restrictions will only apply to the most powerful systems. The text adopts the principles of existing European product safety regulations, which impose primarily company-based controls. At the heart of the project is a list of rules imposed only on systems considered “high risk”, especially those used in sensitive areas such as critical infrastructure, education, human resources, law enforcement, etc.

These systems will be subject to a number of obligations, such as ensuring human control of the machine, creating technical documentation or implementing a risk management system. The legislation provides for special oversight of artificial intelligence systems that interact with humans. It will force them to inform the user that they are in contact with the machine.

Bans will be rare. They will cover applications that conflict with European values, such as citizen rating systems or mass surveillance used in China, or remote biometric identification of people in public places to prevent mass surveillance of the population. On this last point, however, states have obtained exemptions for certain law enforcement missions, such as counterterrorism.

Unlike the voluntary codes of conduct of some countries, European legislation will be equipped with means of supervision and sanctions with the creation of a European Office for Artificial Intelligence within the European Commission. For the most serious offences, it will be able to impose fines of up to 7% of turnover with a lower limit of 35 million euros.

Leave a Comment