UK Treasury brings in Tony Blair's institute to advise on AI in public services

When the government of a major Western country decides to modernise public services with artificial intelligence, one key question arises: whom does it call in as advisers? The UK Treasury has given a revealing answer, and it has not gone down well with everyone.

On Wednesday, Chief Secretary to the Treasury James Murray held a closed-door meeting. Invitees included the AI director of the Tony Blair Institute for Global Change (TBI), the chair of IBM's board of directors, senior executives of Faculty AI, now part of Accenture, and Dex Hunter-Torricke, a former communications adviser at Google, Facebook and Elon Musk's SpaceX. The topic was the strategy for rolling out AI systems across British government bodies.

The Tony Blair Institute has long cast itself as one of the leading advocates of digital transformation in public administration. Founded by the former prime minister, the organisation consistently argues that AI can radically improve public-sector efficiency, from healthcare to tax administration. It publishes large-scale reports, holds conferences and builds bridges between governments and the technology industry worldwide. But it is precisely this dual role, as both an independent think tank and an intermediary for the private sector, that draws sharp criticism.

The reaction from civil society was predictably harsh. Campaigners for digital equality and transparency in technology policy likened the event to "inviting foxes to advise on the future of the henhouse." Behind the vivid metaphor lies a systemic problem: when commercial companies that want to sell AI solutions to the state also shape the strategy for buying and deploying them, a conflict of interest becomes practically unavoidable. Faculty AI, for example, already has a history of major contracts with the British government. The company has worked with the Home Office and other departments, and its merger with the global consulting giant Accenture has only strengthened its position in the public procurement market.

The problem goes far beyond a single meeting. Britain, like many other countries, faces a strategic choice: build its own public-sector AI capability or rely on private-sector expertise. The first path requires large investment in people, infrastructure and research. The second is faster and cheaper at the start, but creates long-term dependence on commercial suppliers whose priorities do not always match the public interest. The presence at the table of former senior executives of the world's biggest technology corporations, from Meta* to SpaceX, makes plain which way the balance is tipping.

For a Russian observer, this situation is of particular interest. Russia is following a similar path in digitising public services, and the role of private technology companies in that process is no less contentious. Britain's experience, both good and bad, will inevitably become a case study. The key lesson is simple: transparent decision-making and a clear line between advisers and beneficiaries of government contracts are not a bureaucratic formality but a fundamental condition for citizens' trust in a digital state.

The broader context is also worth noting. In recent years the Blair Institute has become one of the most influential non-governmental centres shaping the technology agenda for governments in dozens of countries. Its recommendations for bringing AI into public administration sound ambitious: automating routine processes, predictive analytics for allocating resources, personalised public services. All of this sounds appealing on paper. But the devil, as always, is in the details, and above all in who will design, build and maintain these systems while gaining access to the data of millions of citizens.

The British government has not yet commented on the criticism. But the fact that a meeting of this level provoked an immediate and sharp public reaction says a lot. The era when technology decisions for the state were made behind closed doors without public oversight is coming to an end. The only question is whether democratic oversight institutions can adapt to the speed at which AI is entering government, or whether decisions will be made before society has time to make sense of them.

*Meta has been designated an extremist organisation and is banned in Russia.