Europe Reaches Landmark Agreement on AI Regulation: A New Era for Technology Governance

December 2023 saw a significant surge in AI-related acquisitions, as established companies across sectors like biotech, legal, fintech, and edtech moved swiftly to acquire AI startups to enhance and accelerate their own AI strategies. This trend underscores an urgent need for companies to innovate and adapt within a fast-evolving technological realm. Consequently, understanding the implications of these acquisitions goes beyond mere valuation metrics; it invites a closer examination of the regulatory framework surrounding this wave of investment in AI advancements.

To provide some context,  AI aims to create machines that can "think" or act in ways that mimic human capabilities. Under this umbrella term is “generative AI”, such as ChatGPT, which, unlike traditional AI, can produce novel outputs that mimic the patterns and structures of the training data. Since AI’s inception, there has been a growing consensus for the need of regulations to govern AI systems, ensuring they are safe, transparent, and respectful of fundamental rights. In December 2023, the European Union reached a landmark agreement on the AI act, “one of the first comprehensive attempts to globally govern AI, setting a precedent similar to the GDPR for data protection” (Council of the EU, 2023). It is poignant to note that AI has been distinguished from simpler software systems within the act to ensure the right mechanisms are being targeted. The European Parliament “will vote on the AI Act proposals early next year, but any legislation will not take effect until at least 2025” (BBC, 2023).

So, what does the act entail, and what is it trying to achieve? Overall, the AI Act adopts a risk-based approach, imposing stricter rules on AI systems that pose high risks and safeguards on the use of AI within the EU, as well as limitations on its adoption by law enforcement agencies (BBC, 2023). The act aims, per MIT’s technology review, to mitigate harm in areas “where using AI poses the biggest risk to fundamental rights, such as health care, education, border surveillance and public services” (Heikkila, 2023). According to a publication from Baker McKenzie, “obligations are imposed on technology producers and deployers based on the risk category into which their technology fits” (Baker McKenzie 2023). This means that only technologies that pose “unacceptable” levels of danger which will be forbidden, including biometric identification systems and those that use manipulative techniques such as predictive policing systems. Transparency is at the forefront of dealing with foundation models such as ChatGPT, which will be overseen by the European Commission AI Office. Safeguards will focus on stages such as the design and testing of AI mechanisms and will include registering models in an EU database.

What about enforcement? It has been made clear that the EU AI Act does not apply “outside the scope of EU law and will not target AI systems used solely for research or innovation purposes” (Baker McKenzie, 2023). In addition, the AI Act does not apply to AI systems that have been developed exclusively for military and defence uses – National Security will remain untouched (Heikkila, 2023). However, for those caught by the act, the penalties for failing to comply are not light. An article published by Reuters stated that “fines for violations would range from 7.5 million euros ($8.1 million) or 1.5% of a company’s turnover to 35 million euros or 7% of global turnover” (Chee et al, 2023).

The EU AI Act represents a watershed moment in the global regulation of artificial intelligence, charting a path that balances innovation with accountability. Rather than stifling technological progress, this legislation is a crucial step toward fostering an environment where AI can be developed responsibly with the protection of fundamental rights at the core.