In a significant development, the European Parliament has recently given its approval to the Artificial Intelligence (AI) Act, marking a crucial milestone in the regulation of AI in Europe.
By a significant majority, with 499 votes in favor, 28 against, and 93 abstentions, the European Parliament has passed legislation known as the Artificial Intelligence (AI) Act. The Act is designed to ensure that AI systems developed and deployed in Europe comply with EU values and rights, such as human oversight, safety, privacy, transparency, and non-discrimination.
Under the new rules, a risk-based approach will be taken, imposing obligations on both AI providers and users depending on the level of risk associated with the technology. Certain high-risk AI practices, including social scoring and applications that are intrusive or discriminatory, will be prohibited.
The legislation specifically targets AI systems that have the potential to infringe on individuals’ rights and freedoms. This includes systems that can perform real-time or retrospective identification of individuals in public spaces, categorize individuals based on sensitive attributes, engage in predictive policing, utilize emotion recognition in sensitive contexts, or establish facial recognition databases without proper consent.
Additionally, the definition of high-risk AI applications has been broadened to encompass systems that may pose risks to people’s health, safety, fundamental rights, or the environment. The scope also extends to AI employed for voter manipulation and recommender systems on social media platforms with a large user base.
Brando Benifei, co-rapporteur of the draft act, remarked “While Big Tech companies are sounding the alarm over their own creations, Europe has gone ahead and proposed a concrete response to the risks AI is starting to pose.”
Under the approved Act, generative AI systems such as ChatGPT would be obligated to clearly disclose that the content they generate is AI-generated. This measure aims to enable the identification of deepfakes and prevent the dissemination of illegal content. Additionally, these systems would be required to provide comprehensive summaries of the copyrighted data used during their training, ensuring transparency to the public.
To foster innovation and support small businesses, the legislation includes exemptions for research activities and open-source AI components. It also promotes the establishment of regulatory sandboxes, which serve as controlled environments for testing AI applications in real-life situations prior to their deployment. These sandboxes enable developers to assess the performance, risks, and compliance of their AI systems, ensuring a balance between innovation and regulatory oversight.
The AI Act grants citizens the authority to lodge complaints regarding AI systems and ensures they receive explanations for decisions made by high-risk AI systems that have a substantial impact on their fundamental rights. The Act also entails reforms for the EU AI Office, which will be responsible for supervising the implementation of the AI rulebook.
By adopting its negotiating position on the AI Act, the European Parliament showcased its dedication to striking a balance between fostering innovation and safeguarding the rights of citizens within the realm of AI.
Two years ago, the Commission introduced its initial proposal for draft rules with the objective of establishing a worldwide benchmark for AI, a technology of significant importance to nearly every industry and business. The aim was to bridge the gap with AI leaders such as the United States and China.
The next step involves lawmakers engaging in detailed discussions with European Union member states to refine the draft rules before they can be enacted as legislation. A press release indicates that the goal is to achieve a consensus by the end of this year. This process of consultation and negotiation will shape the final version of the AI Act, ensuring that it reflects the collective input and considerations of all relevant stakeholders.
