Global: Agreement Reached on AI Risk Assessment Standards
The Department of Science Innovation and Technology (DSIT) announced that the United Kingdom, along with 26 other nations, signed the Seoul Ministerial Statement. This agreement aims to establish common risk thresholds for the development and deployment of advanced artificial intelligence (AI).
The Statement emphasizes the critical importance of ensuring safety, fostering innovation, and promoting inclusivity in AI development. It outlines a joint commitment to take practical steps to harness AI's benefits while addressing its inherent risks. It also highlights the need to protect against a comprehensive range of AI risks by adhering to principles of transparency, interpretability, explainability, privacy, and accountability.
The Statement calls on all involved parties to enhance accountability and transparency throughout the AI lifecycle by evaluating, preventing, mitigating, and addressing adverse impacts. Additionally, it emphasizes the role of frameworks in managing AI risks and the importance of credible external evaluations and identifying thresholds for severe risks.
The Statement advocates for the application of AI in the public sector and key industries. It promotes making AI resources accessible to small and medium enterprises (SMEs), startups, academia, and individuals while protecting intellectual property rights. Furthermore, the Statement stresses the importance of sustainability in AI development and calls for environmentally friendly practices and workforce upskilling.