The European Union enacted the Artificial Intelligence Act (AI Act) in late 2023, enacted on 1 August 2024, with gradual implementation. Key prohibitions outlined in Article Five, including restrictions on manipulative practices, social scoring, and real-time biometric identification systems, will take effect on 2 February 2025, but there are exceptions.
Jose-Miguel Bello Villarino, a Senior Research Fellow at the University of Sydney Law School, highlights the complexity of the regulation, particularly regarding high-risk systems and how provisions will be monitored. He noted concerns about the challenges of implementing prohibitions, especially around real-time facial recognition, which has potential exceptions in critical situations.
The manipulation of vulnerabilities, such as exploiting gambling habits, raises further questions about enforcement. Bello Villarino advocates for a regulatory framework that prohibits the exploitation of vulnerabilities and emphasises the importance of interoperability for Australia in light of the EU Act. Overall, advanced democracies agree on some degree of AI regulation based on risk.
He noted concerns about the challenges of implementing prohibitions, especially around real-time facial recognition, which has potential exceptions in critical situations. The manipulation of vulnerabilities, such as exploiting gambling habits, raises further questions about enforcement. Bello Villarino advocates for a regulatory framework that prohibits exploitation of vulnerabilities and emphasizes the importance of interoperability for Australia in light of the EU Act. Overall, there is a consensus among advanced democracies for some degree of AI regulation based on risk.
Source: LSJ
Leave a Reply