The European Parliament recently passed groundbreaking legislation aimed at regulating artificial intelligence, after nearly three years of debate and negotiation. The AI Act, which garnered 523 votes in favor and 46 against with 49 abstentions, is designed to protect fundamental rights, democracy, and environmental sustainability while fostering innovation.
Key Provisions of the AI Act
The legislation, which is still pending final approval by the European Council, outlines specific obligations for AI applications based on their potential risks and impacts. Most of the provisions will come into effect 24 months after the law is enacted, with bans on certain high-risk applications taking effect within six months.
Banned Practices
Under the AI Act, the EU is set to ban various practices deemed detrimental to citizens’ rights. Biometric categorization systems based on sensitive characteristics, untargeted scraping of facial images, social scoring, emotion recognition in schools and workplaces, and AI that manipulates human behavior or exploits vulnerabilities are all prohibited. Predictive policing based on profiling individuals or assessing sensitive characteristics will also face restrictions.
High-Risk Applications
Applications deemed high-risk, such as the use of AI in law enforcement and healthcare, must adhere to strict guidelines to prevent discrimination and ensure compliance with privacy regulations. Developers are required to demonstrate transparency, safety, and explainability of the systems. Additionally, AI systems classified as low-risk, such as spam filters, must disclose to users that they are interacting with AI-generated content.
Generative AI and Manipulated Media
The AI Act also addresses concerns related to generative AI and manipulated media. All AI-generated images, videos, and audio, including deepfakes, must be clearly labeled. Furthermore, AI models must respect copyright laws, and rightsholders have the option to reserve their rights to prevent text and data mining unless it is for scientific research purposes. Providers of general-purpose AI models must obtain authorization from rightsholders for text and data mining on reserved works.
In Conclusion
The approval of the AI Act represents a significant step towards regulating artificial intelligence in the EU. By balancing innovation with safeguards for fundamental rights and ethical standards, the legislation is poised to establish Europe as a global leader in AI development and deployment.