Displaying items by tag: AI safety

Representatives and companies from 28 countries, including the USA, China, and the EU, have signed a declaration on AI safety at a summit held at Bletchley Park. This aims to address the risks associated with advanced AI models, particularly ‘frontier AI’ models like those developed by OpenAI, which include large language models like ChatGPT. One notable participant was the billionaire Elon Musk, co-founder of ChatGPT, who regards AI as a risk to humanity: see The declaration seeks to identify shared concerns regarding AI safety risks and develop risk-based policies across countries. The UK government has described this agreement as a ‘world-first’ achievement. However, some experts believe that it lacks concrete policies and accountability mechanisms. They emphasise the need for swift action to establish globally recognised best practices. In addition to the agreement, the UK government announced a £225 million investment in a new AI supercomputer, Isambard-AI, to be built at the University of Bristol. It is expected to be ten times faster than the UK's current fastest machine and aims to make breakthroughs in fusion energy, healthcare, and climate modelling.