UK Prime Minister announced a collaborative arrangement with “like-minded countries and AI companies” to test AI models for safety before to publication, which is a noteworthy step.
UK Prime Minister Rishi Sunak has announced a collaborative arrangement with “like-minded countries and AI companies” to test AI models for safety before to publication, which is a noteworthy step.
In his remarks at the AI Safety Summit, Sunak emphasized that this accord expands upon the G7’s Hiroshima Process and the current “global partnership on AI”.
Under this plan, new AI models will undergo testing by the AI Safety Institute, which succeeds the Frontier AI Taskforce led by Ian Hogarth. The nature of this agreement, whether voluntary or binding, remains unclear.
Sunak assured that safety measures would be enforced “with the public sector capability to test the most advanced frontier models”.
The Prime Minister revealed that AI companies participating in the summit, including OpenAI (developers of ChatGPT) and Elon Musk’s xAI, have granted the UK “privileged access” to their technology.
Sunak outlined plans for the Safety Institute to establish an evaluation process in time to assess the next generation of models before their deployment in the coming year.
The Labour Party earlier declared their intention to implement a binding requirement for AI companies to subject new models to independent safety tests before release if elected.
The AI Safety Summit, held on November 1 and 2, marked the first in a series of international meetings focused on addressing the potential risks associated with AI.
China’s participation in the summit drew attention due to geopolitical tensions and allegations of spying. Despite initial skepticism, Sunak emphasized that both inviting and securing an agreement with China were successful decisions.
In parallel, the US recently announced its own AI safety institute. This development followed President Joe Biden’s executive order mandating that AI developers share safety results with the US government, showcasing a concerted global effort to ensure the responsible development and deployment of AI technology.