British Prime Minister, Rishi Sunak, says an agreement has been reached with like-minded governments and artificial intelligence companies, regarding the rigorous testing of new AI models before their release, to ensure safe development and deployment of AI.
He revealed this during the two-day AI Safety Summit, held at Bletchley Park, UK, on Wednesday and Thursday, 1st and 2nd November. The summit addressed the risks and opportunities associated with AI technology. On Wednesday, the UK secured China’s support for an international effort to manage and mitigate the potential risks posed by AI technology.
Building on this momentum, a “landmark agreement” was announced on the second day of the summit. Prime Minister Sunak expressed the collective commitment of “like-minded governments and AI companies” to collaborate on safety testing for new AI models. The goal is to evaluate AI models to seek out potentially harmful capabilities, and ensure their safety before and after deployment. China was, however, not included in the list of countries participating in this safety testing collaboration.
China not Included in Summit’s Second-Day Talks
While China had been included in discussions during the summit’s first day, its representatives were not included in the second-day talks. Representatives from several countries and regions, including the US, European Union (EU), France, Germany, Italy, Japan, South Korea, Singapore, Canada, and Australia, actively engaged in these discussions.
The importance of testing AI models for safety and potential risks was emphasized in a chair’s statement, summarizing the participants’ discussions. The statement noted, “it is critical that frontier AI is developed safely and that the potential risks of new models are rigorously assessed before and after they are deployed, including by evaluating for potentially harmful capabilities.”
Prime Minister Sunak addressed questions about China absence in the discussions concerning AI model testing. He clarified that the decision to involve China in specific parts of the summit agreed with the UK’s policy to engage with China where possible, while also mitigating risks.
As the AI landscape continues to evolve rapidly, this international agreement represents a significant step toward promoting the responsible and safe development of AI models, ultimately safeguarding against potential risks and harm. It also reflects the willingness of governments and AI companies to work together to shape the future of AI technology in a manner that aligns with global security and safety concerns.