Rishi Sunak has expressed his belief that the UK’s AI safety summit will “tilt the balance in favor of humanity”. This comes after governments and tech companies reached a “significant agreement” to test AI models before their release. While acknowledging that the event at Bletchley Park is just the beginning of the conversation, the prime minister stated that it demonstrates the willingness and capability to control AI technology. Powerful AI models such as OpenAI’s ChatGPT and Google’s Bard rely on extensive data to respond to prompts and make predictions. However, there are concerns about the lack of transparency surrounding the training data. Mr. Sunak emphasized that government regulation will require greater openness about how these models function. The agreement with AI companies to collaborate on safety testing before launching new models is viewed as a crucial step. UK and US governments will establish their own AI safety institutes to conduct such testing and share their findings. However, not everyone at the summit appeared convinced by this arrangement, with Elon Musk seemingly mocking the politicians involved just hours before his meeting with the prime minister. Despite this skepticism, the summit and the resulting Bletchley Declaration have brought together 28 nations, including the US and China, to address safety concerns related to AI models. Mr. Sunak highlighted the potential of AI to transform various sectors but cautioned about the risks it poses, comparing them to pandemics and nuclear war. The Bletchley Declaration emphasizes the need for international cooperation in tackling these threats and outlines plans for future global summits. However, critics argue that more concrete approaches to regulation and addressing present dangers such as job losses and misinformation are necessary. The UK government has already announced that leading AI firms will share their models for research, and it has established a safety institute to identify any concerns. The White House has also introduced similar plans, including the watermarking of AI-generated content to combat deepfake content. US Vice President Kamala Harris, who attended the UK summit, has emphasized the importance of addressing everyday threats alongside future risks. Mr. Sunak has taken a cautious approach to AI safety legislation, fearing that it could hamper innovation. Instead, existing regulators such as the Competition and Markets Authority, Ofcom, and the Health and Safety Executive have been tasked with applying key safety, transparency, and accountability principles to AI technologies.