On 21 July 2019 Boris Johnson visited the home of Dominic Cummings, then best known as the former director of the Vote Leave campaign. Johnson was due to be appointed prime minister in the following days and had a proposition for the political adviser: would he join him in Downing Street and help him to resolve the continuing fall-out of the Brexit referendum?
Cummings told MPs on 17 March that he agreed to the plan on the basis that four conditions were met. “First of all you’re deadly serious about actually getting Brexit done and avoiding a second referendum. Secondly, double the science budget. Third, create some [D]arpa-like entity. And fourth, support me in trying to change how Whitehall and the Cabinet Office work… And he said ‘deal’.”
In last year’s spring Budget, Rishi Sunak announced that one of Cummings’ long-held ambitions would finally be realised. The Chancellor committed £800m to the creation of a British equivalent of the Defense Advanced Research Projects Agency (Darpa), a US institution which played a major role in the early development of the internet and personal computing. But just eight months later, Cummings left Downing Street amid a failed coup. His vision for a British Darpa freed from Whitehall procurement rules, pay structures and investment constraints had yet to materialise.
In a meeting of parliament’s Science and Technology Committee on 17 March, the chair and Conservative MP Greg Clark challenged Cummings on why it was taking so long to set up the agency. After citing Brexit preparations and the pandemic, Cummings spoke of what he sees as a more fundamental issue within Whitehall.
“Doing something like this is so contrary to how the normal system works that it’s very hard to push it through,” said Cummings. “We had the Treasury come back in September-October time last year and say, ‘hang on we don’t want to have primary legislation at all. My god, if we do this and we start enshrining in legislation that some entities don’t have to follow all of our bureaucracy, where will it end? Suddenly everyone will have this wonderful freedom.’ The whole episode shows just how hard it is for Whitehall to deal with the concept of creating something with low friction.”
Despite the Treasury’s reservations, the government introduced a bill to parliament on 2 March that sets out the legal mandate for the “high risk, high reward” Advanced Research and Invention Agency (Aria). Under the plans the agency, which will be exempt from the Freedom of Information Act, will receive £800m in government funding. Although this is just a single-figure percentage of the government’s total annual research and development budget, Aria’s supporters hope that by separating the organisation from Whitehall bureaucracy, it will be able to make longer term and more significant bets on emerging technologies, in the same way that Arpa, Darpa’s predecessor, did in the 1960s.
But Cummings is concerned that the agency will not have enough freedom. Responding to a question about provisions in the bill for the government being able to assume control of the organisation at any time, Cummings warned: “There are too many restrictions – certainly in my model of it you wouldn’t have ministers anywhere near how it spends money. I think that’ll be a disaster. In my model, it would be extremely simple. You would find a director. You would have a maximum of four trustees so that they are actually real trustees and they have real control – not one of these normal government things with 20 people on a board so that no one’s actually exercising serious responsibility for it – and you would cut it loose of the rest of the system. So no I’m not confident about how it will work out.”
Later in the session, the Labour MP Graham Stringer raised “the really difficult balance between accounting for the expenditure of public money and giving scientists the freedom to innovate and think the unthinkable”. Questioning the Business Secretary Kwasi Kwarteng about the plans, Stringer said he was “slightly disappointed when I read the bill because it effectively gives you as secretary of state, or the government, the ability to completely take control. Even if you don’t use [that power], it’s hanging over the organisation at any time. Do you think you’ve got the balance right in giving the necessary freedoms to this body?”
Kwarteng said he was confident that he had, but conceded later that if the corporate governance structure was wrong, he would again be “up here answering questions in front of Mr Clark and yourselves”.
One of the continuing uncertainties surrounding the new agency is where its focus should lie. Cummings suggested that the decision should be left to the experts appointed to lead the agency, rather than to politicians. He ruled himself out of taking on a role, but recommended several scientists he thinks should be involved, including the American physicist Stephen Hsu, who was demoted by Michigan State University in 2020 following an outcry over comments he had made about research into race and eugenics. The physicists Michael Nielsen and Chiara Marletto, and the mathematician Timothy Gowers, were also cited by Cummings as possible candidates.
Speaking later in the meeting, Kwarteng reiterated that the core mission should be decided by “the people we hire”, but added that his department was devising an innovation strategy that will provide a broad focus for British research. “It will be up to the head of Aria to decide whether he or she should adopt what the innovation strategy suggests, adapt it or reject it, but I think there is a role for government to stress or outline tramlines where the UK has comparative advantage and what areas we want to see dynamic innovation in.”
The government is expected to launch the search for Aria’s first chief executive in late spring.