On Thursday 11 May, the European Parliament took a decisive step towards the regulation of artificial intelligence (AI). Preliminary approval of the EU AI Bill represents a significant step forward in EU efforts to limit potential abuses of AI, while continuing to foster innovation in this growing field.
A long journey
The regulatory proposal was first presented by the European Commission in Brussels two years ago. However, its consideration has been delayed due to controversies around the potential dangers of generative AIs, systems capable of producing textual content and images autonomously. It was not until the end of 2022 that EU Member States managed to define their position on this complex issue.
The challenges of AI
AI systems arouse both fascination and concern. On the one hand, they have the potential to save lives by enabling major advances in the field of medical diagnostics. On the other hand, they can be used for mass surveillance purposes by authoritarian regimes. Apps like OpenAI’s ChatGPT textual content generator have shown the immense possibilities of AI, while the dissemination of fake images created by apps like Midjourney have alerted to the risks of manipulation of public opinion.
A business-based approach
The European Parliament has confirmed its support for the approach proposed by the Commission, which draws on existing regulations on product safety. This approach imposes controls that rely primarily on the companies that develop and implement the AI systems. At the heart of the project is a list of rules imposed on applications deemed “high risk”, which would include all systems used in sensitive areas such as critical infrastructure, education, human resources, law enforcement. and migration management.
MEPs take a stand
MEPs have also expressed their desire to set up a specific regime of obligations for generative AIs, such as ChatGPT. They want to force providers to put in place protections against illegal content and to reveal the data used to develop their algorithms. The Commission’s proposal, presented in April 2021, already provides a framework for AI systems that interact with humans, requiring them to inform the user when interacting with a machine.
In addition, MEPs are considering adding a ban on emotion recognition systems, removing derogations allowing remote biometric identification in public places by law enforcement, and banning the bulk collection of photos on the Internet to train the algorithms without the consent of the persons concerned.
Regulation based on European values
Bans will not be numerous, but they will be targeted. Applications that run counter to European values, such as the citizen rating systems or mass surveillance used in China, will in particular be targeted by these bans.
The regulation proposed by the EU is unprecedented and puts Europe at the forefront of international efforts to control the societal impacts of AI. As technology continues to evolve at a rapid pace, the need for proper and ethical regulation is more pressing than ever.
Towards a confirmation in June
MEPs voted for their position, which will have to be confirmed in plenary in June. This step will mark the start of the difficult negotiations between the different EU institutions to reach an agreement on the final text of the regulations.
As Europe strives to chart a balanced course between technological innovation and ethical protections, other regions of the world will watch closely. Decisions made in the coming months may well shape the future of AI globally.