OpenAI signs agreement to help modernize UK government services

OpenAI has taken an important step to advance its global ambitions by signing a strategic agreement with the UK government to deploy its AI models and background expertise to overhaul government services.

With high and unsustainable UK public spending coupled to a hidebound central government, government departments need to address weak delivery and chronic public sector waste. The ambition is that AI will help them do so.

There’s a lot at stake for both sides in this deal. The UK government will need to see results at some point, while OpenAI will see the challenge as a chance to prove its technology in the face of some skepticism.

The agreement doesn’t go into much detail about how OpenAI’s technology will be deployed, but it’s clear that the government and the Department for Science, Innovation and Technology (DSIT) see the US company as an important partner in their quest to evaluate the technology’s potential.

The focus will be on helping the government implement the priorities set out in January’s AI Opportunities Action Plan. Beyond that, the government wants to embed AI knowhow in the wider economy while encouraging companies in the sector, including OpenAI itself, to expand their footprint in the country.

“AI will be fundamental in driving the change we need to see across the country — whether that’s in fixing the NHS [National Health Service], breaking down barriers to opportunity, or driving economic growth,” said UK Technology Secretary Peter Kyle. “This partnership will see more of their work taking place in the UK.”

OpenAI CEO Sam Altman was also enthusiastic. “Britain has a strong legacy of scientific leadership, and its government was one of the first to recognize the potential of AI through its AI Opportunities Action Plan. Now, it’s time to deliver on the plan’s goals by turning ambition to action and delivering prosperity for all,” he said.

The government announcement alluded to the need to build a domestic AI capability and not simply offshore expertise in a way that might lead to technology dependency. The agreement would, “support the UK’s goal to build sovereign AI in the UK: ensuring that the UK continues to drive critical AI research and participates actively in development of this unique technology,” said the official Memorandum of Understanding (MoU).

Well-worn path

In the UK, public sector IT projects have a habit of going off the rails or costing too much. The most extreme example was the infamous Post Office Horizon scandal that led, over more than a decade, to the wrongful prosecution of hundreds of innocent users for accounting fraud.

The worry is that AI will eventually follow this well-worn path of unfulfilled potential. In May, a Parliamentary Public Accounts Committee (PAC) report found that, so far at least, AI’s benefits inside government have been modest.

However, the announcement with OpenAI hints at more realistic expectations. Initially, the government will probably be happy if they can encourage OpenAI to expand its local AI infrastructure and workforce.

This is already bearing fruit. “Alongside the partnership agreement with the UK Government, OpenAI confirmed its intention to increase its footprint in the UK,” said OpenAI in its announcement of the partnership.  “OpenAI will share more details on its expansion in the summer.” Its first European office opened in 2023 in London and currently employs around 100 people.

‘Big AI’ risk

But OpenAI’s motivation might not be just all about billing for project work. As a government partner, it will also get access to a lot of data, an essential ingredient in improving AI’s usefulness.

However, the advent of AI in government is raising concerns among some UK businesses.

“Experience shows that introducing new tools into existing systems without proper restructuring typically creates expensive chaos,” said Karli Kalpala, head of strategic transformation for UK and Ireland at business automation company Digital Workforce. “Departments need to ensure that they do not deliver AI in siloes, which could create disconnected initiatives that duplicate effort whilst solving little of the substance.”

“If ministries use AI merely to accelerate current processes, they’ll create sophisticated digital assistants rather than intelligent, autonomous systems and workers that can deliver genuine transformation,” Kalpala said, noting that getting into bed with big AI would never be risk-free.

“Single vendor partnerships create dangerous dependencies,” he pointed out. “A more sensible approach would involve specialist AI companies with deep sector knowledge, such as healthcare AI firms for the NHS or defense technology specialists for security services. These focused partnerships would reduce vendor lock-in risks.”

Hilary Stephenson, managing director of usability design company, Nexer Digital, also has doubts. To fulfil its ambitions, she argued, the government will need to develop inhouse AI knowledge and not simply outsource all the thinking to external partners.

Its mistake was to see AI as purely about technology, she said. “We welcome innovation, but with it comes a responsibility to maintain transparency, protect public trust, and avoid deepening existing inequalities. People have a right to know how their data is being used.”