UK government yet to trial OpenAI tech months after signing partnership | Technology policy
When the UK government signed a memorandum of understanding with OpenAI, the technology firm behind ChatGPT, the partnership was hailed as one that could leverage artificial intelligence to “tackle society’s biggest challenges”.
But eight months after the fanfare of that announcement, the government has yet to conduct any trials of the company’s technology.
A freedom of information (FoI) request asked the Department of Science, Innovation and Technology (DSIT) for information about the trials being carried out under the memorandum, which said the company would work with civil servants to “identify opportunities for how advanced AI models can be applied across government and the private sector”.
The ministry said it had none of this information and “has not conducted any trials within the scope of the memorandum of understanding with OpenAI.”
In response to a question from the Guardian, DSIT pointed to an agreement signed by the Ministry of Justice (MoJ) last October. enabled civil servants to use ChatGPT “with the option of UK-based data storage for customers”.
Tarek Nseir, CEO of AI consultancy Valliance, which offers FoI, said: “Either there was a massive failure in execution or, in my view, it was a mistake of intent.
“There are undoubtedly parts of the government that are interested in these leading models and these providers… We have little to show for it.
“The introduction of ChatGPT in one department does not quite reflect the intent of the MoU.”
He added: “We use PowerPoint; that does not mean we have a strategic relationship with Microsoft. If that is the intention of the MoU, then our government is not taking seriously the impact of AI on our economy.”
The Justice Department’s agreement to use ChatGPT appeared to be part of a larger agreement.Artificial Intelligence Action Plan for JusticeIt was launched separately last July. DSIT also said it continues to work with the UK AI Security Institute to test AI models and develop security measures in collaboration with OpenAI.
The statement said: “We are pleased with the progress we have made on the memorandum of understanding with OpenAI. This work is active, ongoing and focused on delivering real results for public services and the economy.”
The department also said it will work with Nvidia and Nscale to “deploy GPUs for Stargate UK, with a focus on strengthening the UK’s AI capabilities.”
None of this — other than ChatGPT at the Department of Justice — amounted to a government-wide deployment of advanced AI models as described.
OpenAI said the scope of the FoI did not cover the full extent of its activities in the UK and that it was “proud of the progress we have made on the memorandum of understanding with the UK government”.
An investigation by the Guardian has found that although Nscale has promised to build the UK’s largest supercomputer using Nvidia’s GPUs by the end of 2026, it will almost certainly not be able to complete the project on time and has publicly misrepresented its progress on the site.
Nscale will also collaborate with OpenAI on Stargate UK; this initiative will attempt to deploy potentially 8,000 Nvidia chips to sites across the UK; however, the exact language of the press release was unclear.
Contacted by the Guardian, OpenAI said it had “nothing to share” about the progress of this deployment, which it had previously suggested would happen this quarter.
The government’s memorandum with OpenAI was one of a series of high-profile deals outlining how AI could transform “the way people live, learn, work and access public services” and “can be a powerful tool to increase productivity, accelerate discovery and create opportunity”.
Matt Davies, economic and social policy lead at the Ada Lovelace Institute, said: “AI could transform how people interact with public services, but government experiments with these technologies must be open and transparent. “Voluntary partnerships with major AI companies do not follow the usual procurement rules, which raises real questions about accountability and scrutiny.
“The memorandum with OpenAI does not clearly explain how progress will be measured or how the public benefit will be delivered, and the risks of ‘lock-in’ becoming dependent on a company’s products and services are nowhere addressed.
“The public is concerned about the government’s approach to AI. In our survey, 84% of respondents said they were concerned that the government was putting industry interests ahead of protecting the public. The government needs a positive vision of how AI can truly improve people’s lives; just aiming for ‘more AI’ is not good enough.”
The government signed similar agreements with Türkiye. anthropic, Google DeepMind And Nvidia. The Guardian understands the Google memorandum, signed in December, is in the early stages of planning.
Anthropic said it plans to create an AI assistant to help navigate government services and is also working with the UK AI Security Institute to conduct security research.
Nvidia did not respond to a request for comment.
