A groundbreaking judgment has been declared as the European Union Parliament officially approved the AI Act on June 14, 2023. This legislation represents a significant step towards regulating and mitigating the risks associated with the use of AI. Being one of the pioneering global regulations on AI, this Act establishes strict guidelines on the permissible applications of this advancing technology.
The approved legislation primarily categorizes AI risks into three distinct groups. Firstly, there are AI systems deemed to pose “unacceptable risks,” which the Act explicitly prohibits. Notable examples include social scoring and predictive policing practices observed in China. Secondly, the Act imposes specific limitations on “high-risk” systems, particularly those that could potentially impact voting behaviors. Meanwhile, other AI applications remain unregulated for the time being.
In essence, the Act adopts a risk-based approach to regulate AI. It mandates that AI developers undertake a risk assessment or analysis prior to implementing AI for any intended purposes. However, there remains a lack of clear definition regarding what constitutes risks in the context of AI. Bridging this clarity gap is of utmost importance to avoid impeding innovation under the guise of regulations.
The passing of this Act is being regarded as a landmark development that has the potential to serve as a pivotal model for other nations and regulatory entities worldwide. Its inception stems from the recognition that AI has a pervasive impact on various aspects of individuals’ lives, profoundly influencing their decision-making processes. As a result, the Act aims to establish protective measures to address the unparalleled power wielded by AI and to regulate its utilization by companies. By implementing this legislation, the European Union seeks to set clear boundaries and guidelines to govern the responsible and ethical deployment of AI technology.
Although the Act had been under development for a considerable period, its acceleration towards implementation was primarily driven by the rapid proliferation of AI, particularly with the emergence of advanced systems like ChatGPT. In addition to restricting the use of facial recognition applications, the Act also imposes obligations on developers of chatbot technologies such as ChatGPT, particularly regarding the data utilized in creating their programs. Notably, the latest versions of the Act place significant emphasis on enhancing transparency within the realm of generative AI.
On the other side of the spectrum, the tech giants, including the creator of Open AI are rallying towards a different focus. Although they acknowledge the necessity of AI regulation, they find it challenging to comply with the specific contents and provisions outlined in the EU AI Act. In fact, the final decision on the Act might even require ChatGPT to be taken off the E.U. markets.
As AI continues to evolve at an unprecedented pace and manifest in unexpected ways, the ultimate efficacy of these regulations remains uncertain. The dynamic nature of AI technology presents challenges in accurately predicting its future developments. However, it is imperative to establish regulations to mitigate potential risks and ensure responsible AI deployment. By embracing regulation, the aim is not to stifle innovation but rather to strike a balance between advancement and safeguarding against potential harms.
Following the passing of the Act in the E.U. Parliament, the final version of the law will be put for negotiations and voting between representatives of the three branches of the European Union — the European Parliament, European Commission and the Council of the European Union. The objective is to reach a final decision of the state of AI regulation by the end of 2023.
The passing of this Act has raised an important question regarding its potential to become a benchmark for AI regulation, much like how the EU GDPR law set the standard for data privacy. As AI adoption continues to surge globally, particularly in Asia, it remains uncertain whether regulatory measures introduced in Asian countries will bear resemblance to or draw inspiration from the EU Act. The extent of alignment between different regions’ AI regulations is yet to be determined, and whether the EU Act will serve as a guiding influence remains an open question.