The UK government’s ambition to harness the power of artificial intelligence for public good appears to be lagging, with no trials yet undertaken using technology from OpenAI, the company behind ChatGPT, nearly eight months after a high-profile partnership was announced. The memorandum of understanding (MoU) signed with OpenAI in April aimed to explore how advanced AI models could be deployed across government and the private sector to tackle pressing societal challenges. However, a recent freedom of information (FoI) request revealed a surprising lack of progress, raising questions about the government’s strategy for AI adoption and its ability to translate ambitious goals into tangible results.
The Department for Science, Innovation and Technology (DSIT) confirmed it had not conducted any trials under the MoU with OpenAI. This revelation comes as the UK government increasingly emphasizes its commitment to becoming a global leader in AI, investing heavily in research and development and outlining a vision for the technology to transform public services. The initial fanfare surrounding the OpenAI partnership suggested a swift move towards practical application, but the reality appears to be far more cautious.
Ministry of Justice’s Limited ChatGPT Access
Responding to the FoI request, DSIT pointed to an agreement with the Ministry of Justice (MoJ) last October as the primary example of collaboration. The MoJ enabled civil servants to use ChatGPT, with the option of UK-based data storage. This move, part of a broader “AI Action Plan for Justice” launched in July, allows access to the chatbot for tasks like summarizing legal documents and drafting correspondence. However, critics argue that this limited implementation falls far short of the MoU’s stated ambition.
Tarek Nseir, founder of the AI consultancy Valliance, which filed the FoI request, expressed skepticism. “Either there’s been a huge failure in execution, or it was a failure of intent in my view,” he said. “Rolling out ChatGPT in a department hardly reflects the ambition of the MoU.” Nseir drew a parallel to basic software adoption, stating, “We use PowerPoint – that doesn’t mean we have a strategic relationship with Microsoft. If this was the intent of the MoU then our government is not taking the impact of AI on our economy seriously.”
Broader AI Partnerships and Stargate UK
DSIT also highlighted ongoing work with the UK AI Safety Institute, collaborating with OpenAI to test AI models and develop safeguards. The department pointed to partnerships with Nvidia and Nscale to deploy GPUs for “Stargate UK,” an initiative aimed at strengthening the UK’s AI capabilities. However, details surrounding Stargate UK remain somewhat opaque. The project intends to potentially deploy 8,000 Nvidia chips across the UK, but OpenAI, when contacted, stated it had “nothing to share” on the deployment’s progress, despite previously suggesting it would occur this quarter.
A recent investigation by The Guardian revealed concerns about the timeline and transparency of the Nscale project, suggesting the company has misrepresented its progress towards building the UK’s largest supercomputer, originally slated for completion by the end of 2026.
Accountability and Public Trust Concerns
The lack of demonstrable progress raises broader questions about accountability and transparency in the government’s approach to AI. Matt Davies, economic and social policy lead at the Ada Lovelace Institute, emphasized the need for open experimentation. “AI could transform how people interact with public services, but government experimentation with these technologies must be open and transparent,” Davies said. He also noted that voluntary partnerships with considerable AI companies bypass traditional procurement rules, “raising real questions about accountability, and scrutiny.”
Davies further pointed out that the MoU with OpenAI lacks clear metrics for measuring progress and delivering public benefit, and doesn’t address the potential risks of becoming overly reliant on a single company’s technology – a phenomenon known as “lock-in.” Public concern is also high, with 84% of people polled by the Ada Lovelace Institute expressing concern that the government prioritizes the interests of the AI sector over public protection.
Expanding AI Partnerships
The UK government has signed similar MoUs with other leading AI companies, including Anthropic, Google DeepMind, and Nvidia. The agreement with Google DeepMind, finalized in December, is reportedly still in its early planning stages. Anthropic is developing an AI assistant to help navigate government services and is collaborating with the UK AI Safety Institute on safety research. Nvidia did not respond to a request for comment.
The slow pace of implementation, coupled with a lack of transparency, underscores the challenges of integrating cutting-edge AI technologies into the complex machinery of government. While the UK government continues to articulate its commitment to AI innovation, translating that commitment into concrete action remains a significant hurdle.
The next update on the UK government’s AI strategy is expected in early 2024, with a planned report outlining progress on the various MoUs and initiatives. The report will be closely watched by industry experts and the public alike, as the UK seeks to establish itself as a leader in the rapidly evolving field of artificial intelligence.
What are your thoughts on the UK government’s AI strategy? Share your comments below and let us know how you think AI should be used to improve public services.
