The European Union has taken a step towards regulating artificial intelligence with the approval of its much-awaited AI Act.
The new legislation, which aims to restrict the public use of AI applications considered high-risk, such as deepfakes and facial recognition software, will apply to all companies deploying such technologies in the 27 EU-member states.
Under the AI Act, companies that breach the new rules will face hefty fines of up to 35 million euros or 7% of their annual global revenue.
The Act outlines different risk categories for AI use, ranging from “low-risk” to “high” and “unacceptable risk,” based on the potential harm to consumers. AI applications that threaten individual rights, such as facial recognition software in public places and social scoring, will be banned outright.
While the AI Act aims to protect personal information and ensure trust, transparency, and accountability when dealing with new technologies, there are concerns among startup founders that the law could stifle investment and innovation.
Some worry that the measures could put Europe further behind the United States and China in the AI race, as smaller companies may find it challenging to comply with the new regulations.
To address these concerns and boost innovation, the EU has announced plans to build “AI Factories” that will provide the necessary infrastructure for startups to buy and upgrade. The EU will also grant privileged access to supercomputers for European startups developing “trustworthy” AI that aligns with EU values and rules.
In addition to the AI Act, the European Blockchain Observatory and Forum (EUBOF) has advised the EU Commission to prepare for the convergence of blockchain technology and AI.
The EUBOF report highlights the potential for blockchain to securely store sensitive AI data sets, particularly in healthcare and finance, and enable decentralized AI networks that can reduce the risk of data monopolies and promote collaborative AI development.
