The UK’s Advanced Research and Inventions Agency (ARIA) is three years into its ten-year mandate under law and already has a new chief executive.

The agency is designed to support high-risk, high-reward civilian research and is inspired by the United States’ Defense Advanced Research Projects Agency (DARPA), which has helped pioneer technologies such as the Internet and personal computers.

ARIA’s aim to fund breakthrough scientific breakthroughs began in 2023 with a four-year budget of £800 million ($1.1 billion), of which around 20% funds organizations based outside the UK. Last year, the UK government pledged to increase the agency’s spending to £400 million a year by 2030 to expand ARIA. Programs funded so far include the development of minimally invasive neurotechnologies, ways to allow AI agents to safely interact with each other, and high-altitude floating “cell towers” for communications. It is about to start recruiting its third wave of program directors, each of whom will work on a specific target for up to five years and hand out a budget of around £50m.

Kathleen Fisher, a computer scientist and former director of DARPA’s cybersecurity-focused Information Innovation Office in Arlington, Virginia, took over as executive director of ARIA in February.

Nature He spoke to Fisher at the agency’s London office, in the heart of the King’s Cross tech district, to find out what makes ARIA different, how long it will take to prove itself and whether AI will change science.

Why did ARIA change CEO so soon?

One of the “secret sauce” elements coming out of DARPA is term limits for decision makers. [Fast] Rotation helps because if you make high-risk decisions and you see that not all of them work out, you can start to become much more risk-averse. But if the result of those high-risk decisions ends up on a new team, you don’t have the feeling of “that was my baby and it didn’t work out.”

It also means that you don’t become obsessed with certain ideas. DARPA has invested in AI for 60 years, but that’s not because one person said “AI is the important thing.” Probably hundreds of people said, “No, AI isn’t there yet, but it’s still a good investment.” It is necessary to make deep commitments for short periods of time and then look at them with fresh eyes, again and again.

What is it like working with ARIA funding?

Each program tries to achieve a specific thing; create treatments using minimally invasive deep brain stimulation, for example. This often involves researchers from many different disciplines coming together. We have assessment teams working to measure how ‘creators’ (ARIA’s term for researchers) are doing, entrepreneurs in residence who will help creators with their translation plan, and other partners who could help creators navigate regulatory pathways. There are many intervention points where we evaluate how the teams are doing. Teams that do well can get additional funding, those that don’t can be eliminated.

DARPA ‘doubles’ must make their dreams come true

What characterizes a good ARIA researcher?

Researchers who think the way we work is exciting and empowering are the ideal creators of ARIA. People who say, “Would you leave me alone so I can investigate?” they are not. Curiosity-driven research is vitally important and we need those types of researchers. But they are not the type of investigators who will do well at our agency.

TO Guardian The article highlighted concerns about ARIA cash going to US technology and venture capital companies. Does ARIA spending need more scrutiny?

ARIA’s mission is daunting, we are supposed to change the world. We need to be given the freedom to do so. Accountability mechanisms exist, we answer to the UK National Audit Office and parliamentary select committees. We do our best to be transparent and engaged with the relevant people, so I don’t think we need further scrutiny.

It is also unrealistic to expect the UK to have 100% of the talent needed to achieve our goals, and every time we fund an organization that is not in the UK, we do specific analysis to explain why that organization is suitable and necessary.