The European Union Parliament approved the world’s first comprehensive regulatory framework for artificial intelligence (AI) Wednesday.
This landmark legislation, known as the EU AI Act, categorizes AI technologies into four levels of risk — unacceptable, high, medium, and low — and sets out to ban technologies that pose unacceptable risks, according to a report published by CNBC.
By officially establishing new regulations, the act aims to balance the promotion of AI innovation with protecting fundamental rights in modern society. One pressing example would be AI policy established to protect free and open elections by addressing the potential misuse of AI deepfakes in elections.
According to a report published by The Associated Press, most AI systems are expected to fall into the low-risk category. Examples of these kinds of programs would be spam filters and content recommendation systems. However, it’s reported that high-risk AI tooling, which includes medical devices and AI systems linked to critical infrastructure (i.e. electrical grids or water systems), will face more demanding requirements.
The approval of the AI legislation arrives five years after the regulations were first proposed. The vote also follows the EU’s pattern of more stringent tech regulation, as seen with the General Data Protection Regulation (GDPR) which established certain protections for privacy and security of individual’s personal data. The EU’s AI policy is anticipated by many experts to influence AI policy development in other governments around the world.
The policy will require adjustments in how companies develop, deploy, and manage AI systems. In the workforce this could lead to increased investment in compliance and ethical practices, likely fueling upskilling efforts and training programs to help the workforce navigate new guardrails established by AI regulations and other ethical considerations. The heightened focus on protecting fundamental rights amid fast technological development could lead to new job roles focused on AI ethics, compliance, and governance.
The EU AI Act represents a major step in regulating AI software and tools, and it aims to promote technological progress, but not at the expense of ethical standards. The legislation is expected to come into effect at the end of May 2024, with a staggered implementation beginning in 2025.
