UK Pledges £600 Million (About $800 Million) to Shield Health and Infrastructure From AI Threats

Britain plans to spend about £600 million (around $800 million) to protect people and important services from dangers linked to artificial intelligence (AI). Most of the money will go toward a new effort to protect people from biological threats, such as dangerous germs and diseases.

The government announced the plan on Saturday. It includes £100 million (about $133.7 million) to create a laboratory that will use AI to find biological threats. Another £65 million (about $86.9 million) will help develop ways to fight dangerous germs that cause disease. Britain's AI minister, Kanishka Narayan, said he wants the new AI Biosecurity Lab to become one of the world's leading centers in this field. He said it would bring together top scientists from around the world to protect people from new biological dangers.

Two American AI companies, OpenAI and Anthropic, will provide up to £200 million (about $267.5 million) in credits that let the project use their AI models. Google DeepMind and Basecamp Research, an AI startup based in London that works with life sciences, will also help.

The money will support several areas, including safety, research and British industry. Britain's existing AI Security Institute, which studies AI risks, will get an extra £120 million (about $160.5 million). Another £130 million (about $173.9 million) will support new projects to make computer systems harder to attack. These projects will protect important services and the wider economy. They will operate through the National Cyber Shield and the National Exposure and Real-Time Vulnerability Centre, which focuses on finding weaknesses in computer systems.

Note: The £165 million for biological safety includes £100 million for the laboratory and £65 million for ways to fight dangerous germs. The credits worth up to £200 million (about $267.5 million) from OpenAI and Anthropic are not cash, so they are not included in the spending totals. The National Cyber Security Centre will also receive money to expand its work in Manchester, but the government has not said how much.

The plan will also help British businesses and technology industries. A £60 million fund (about $80.2 million) will invest in British companies using AI to build robots, machines that can work on their own, and manufacturing technology. Another £60 million will pay for AI research and contracts to buy AI products or services. The government will spend £23 million (about $30.8 million) to study how AI affects the economy. It will use £18 million (about $24.1 million) to set up a school for designing computer chips. Another £15 million (about $20.1 million) will pay for a prize that rewards AI solutions to problems governments need to solve, starting with healthcare. Mayors in different regions will share £30 million (about $40.1 million) to support local AI projects.

The announcement comes as governments and security agencies face quickly changing dangers from AI. AI is now being used to find weaknesses in computer software on a large scale. Anthropic said this month that its AI models helped find more than 6,000 security weaknesses in 591 open-source software projects, whose code is available for others to examine and use. These included problems in OpenBSD and FreeBSD that had gone unnoticed for 27 years and 17 years, respectively. Anthropic also reported spending more than $12 billion on running its business last year. This shows how expensive it can be to compete in developing the most advanced AI systems.

This situation has led to a new word: "bugmageddon." It combines "bug," meaning a problem in software, with "Armageddon," meaning a huge disaster. The word describes a possible situation in which AI finds software weaknesses much faster than companies and programmers can fix them. Security researchers have shown that AI can sometimes finish in days work that once took teams of highly skilled experts months. This makes it easier for both attackers to find weaknesses and defenders to protect against them.

Britain hopes that government spending will encourage private AI companies to use their technology to protect people and important systems. The new biological safety laboratory will bring together advanced AI models and the science of how diseases spread. The money for computer security will focus on weaknesses in important services such as electricity and water supplies, transportation and financial systems. These services help keep the wider economy running. The AI Security Institute, which was created to study the dangers of advanced AI models, will get more resources to continue and expand its work.

The credits from OpenAI and Anthropic will let the biological safety program use computing power and run AI models without paying cash for those services. This means the government can make its own money go further. Google DeepMind's involvement brings a third major AI laboratory into the project. Basecamp Research, a British startup, will also take part.

For investors, the plan suggests that the British government will offer a more regular supply of contracts in AI security, biological safety, computer chip design and robotics. British startups in these fields have had trouble growing as large as American competitors that have more money. The regional funding and money given to mayors could also create chances for local companies outside London to win government contracts.

The government's planned spending is small compared with the huge amounts of private money being invested in AI. Officials have not explained when the money will be spent or which government departments will be responsible for it. Whether the new laboratory becomes the world-leading center Narayan hopes for will depend on whether it can hire the right experts. It will also depend on how much access the participating AI companies give government scientists to their models. Even with these unanswered questions, Britain's approach is clear: the government sees AI as something that can create dangers, but also as a tool that can help protect people from those dangers.