UK renames AI Security Institute, drops

TechRadar reports that the United Kingdom government has announced that its AI Safety Institute will now carry the name AI Security Institute. The body itself continues to exist, but the article presents the new title as a clear move toward cybersecurity. It adds that the Labour government, in office since 2024, has committed heavily to AI and that the institute had supported it in that effort.

The piece explains the distinction as it sees it. Security concerns defending against threats such as cyberattacks, and limiting the ways AI could be turned against national security or used to enable crimes including fraud and the development of chemical weapons. Safety, in the article's account, covers a wider field: it also addresses misinformation spread by chatbots and looks at the broader societal effects of AI models, beyond immediate cyber threats.

The article notes that the government's Plan for Change, published in January 2025, relied heavily on AI, with measures such as Growth Zones, the use of public data to train models and a goal of embedding AI deeply in public services. TechCrunch, it says, observed that the document never used the words harm, safety or threat. Science, Innovation and Technology Secretary Peter Kyle is quoted as saying the institute's work will not change, but that the sharper focus will protect citizens and allies from those who would use AI against institutions, democratic values and the way of life.

As part of the same plan, the government has agreed a partnership with the AI company Anthropic. It is intended to realise the opportunities of the technology, with continued attention to responsible development and deployment, and to provide insights on improving public services and advancing science. TechRadar links this to the UK's wish to attract technology investment.