The European Union legislators in the European Parliament have given their preliminary approval for the world’s first artificial intelligence (AI) legislation, known as the AI Act.
The endorsement, following the approval of EU member states, sets the stage for a parliamentary vote scheduled for April.
During the EU meeting, a massive vote was given towards endorsing the provisional agreement on the AI Act.
This historic legislation aims to establish comprehensive guidelines for the use of AI across industries, including banking, automotive, electronics, aviation, security, and law enforcement.
The world is grappling with various challenges in artificial intelligence (AI), including ethical concerns, regulatory hurdles, and impacts on human well-being.
Similar to the crypto industry, there have been concerns about how to create guidelines that can help players safely operate in the industry and the EU is one region that has been at the forefront of this movement.
The United States and the United Kingdom have demonstrated significant interest in this matter. In the U.S., the President’s Office has been engaging in discussions with top tech CEOs to formulate a roadmap for the responsible use of AI.
Meanwhile, the UK hosted the AI Summit last year, bringing together various nations to establish fundamental principles for AI development, aimed at benefiting all involved.
According to the EU, the AI Act specifically addresses foundational models and generative AI, such as OpenAI’s ChatGPT, which are trained on extensive datasets.
The legislation introduces safeguards, including copyright protection for creators, and prohibits AI applications that threaten citizens’ rights, including biometric categorization and social scoring.
After addressing concerns from EU member states, especially France, the endorsement for the AI Act was secured.
The changes focused on easing administrative burdens for high-risk AI systems and improving protection for business secrets.
Following a political agreement in December 2023, efforts began to finalize the compromise text for approval.
The “coreper” vote on Feb. 2, involving representatives from all member states, paved the way for endorsement by European Parliament committees.
The AI Act is now scheduled for a vote in March or April in the European Parliament, and if approved, it is expected to be fully implemented 24 months after taking effect, with some provisions kicking in earlier.
The Artificial Intelligence Act imposes bans on risky AI applications, including biometric categorization and emotion recognition. It adds that law enforcement can use biometrics in specific cases with strict safeguards.
It also outlined transparency requirements for AI training and additional measures for powerful models.
The Act also supports innovation and SMEs with regulatory sandboxes for testing new AI before market placement.
Finally, it awaits a formal adoption in an upcoming Parliament plenary session and final Council endorsement.
