UK yet to trial OpenAI technology months after landmark agreement

Lack of visible progress risks undermining confidence in government strategy

The government has yet to begin testing OpenAI tech, despite ministers previously presenting the partnership as a vital plank of public service reform.

A freedom of information (FoI) request has revealed that, eight months after the signing of a memorandum of understanding (MoU) with the company behind ChatGPT, no formal trials involving its technology have taken place across government.

The agreement, announced in July 2025, was described at the time as a major step towards using advanced AI systems to "address society's greatest challenges" and improve how public services are delivered.

However, in response to a request submitted by consultancy Valliance, the Department for Science, Innovation and Technology (DSIT) confirmed it held no records of any such trials.

The department stated it had "not undertaken any trials under the memorandum of understanding with OpenAI".

The findings raise questions about the pace at which the government is translating its AI ambitions into practical implementation.

Valliance had sought clarity on when the public might begin to see tangible benefits from the partnership, what problems had been prioritised, and how far any pilot projects had progressed.

While no formal trials have been conducted under the MoU, DSIT pointed to a separate initiative within the Ministry of Justice (MoJ).

Since October 2025, civil servants there have been allowed to use ChatGPT, with an option for UK-based data storage. That deployment forms part of a broader "AI Action Plan for Justice", launched independently of the OpenAI agreement.

Beyond this, DSIT highlighted ongoing collaboration with the UK AI Safety Institute to test models and develop safeguards, as well as parallel efforts involving Nvidia and Nscale to expand the country's computing infrastructure.

But critics argue these initiatives fall short of the government's original pledge to deploy advanced AI models across public services.

Concerns over delivery

Tarek Nseir, co-founder of Valliance, said the lack of visible progress risked undermining confidence in both government strategy and the wider adoption of AI.

"There's no shortage of ambition or spending in the UK, but in truth the issue is in delivery," he said.

"If the government's own flagship partnerships aren't producing visible progress, public trust will erode and businesses will hesitate."

He added that uncertainty could contribute to growing scepticism about whether AI can deliver meaningful improvements.

"The future of the country depends on our ability to take the lead in implementing and extracting value from the technology. But without tangible outcomes to show the way forward, the UK risks slipping further behind," Nseir noted.

In a statement to The Guardian, DSIT said work under the agreement was ongoing.

OpenAI also defended the partnership, stating that the FoI request did not reflect the full extent of its work in the UK and that it was "proud of the progress" made so far.

The OpenAI agreement is one of several high-profile deals the UK government has struck with leading AI firms, including Anthropic, Nvidia and Google DeepMind.

However, progress appears uneven.

The agreement with Google DeepMind, signed in December 2025, is understood to still be in early planning stages, while Anthropic has said it is developing an AI assistant to help citizens navigate government services.

A separate investigation found that plans to build a major supercomputer using Nvidia technology are unlikely to be completed on schedule.