Britain devises AI technology to assist the common public in comprehending their puzzling public administration procedures
In the heart of London, the UK government is making significant strides in leveraging Artificial Intelligence (AI) technology. The AI Exemplar program, a series of AI projects aimed at delivering billions in value, is underway. One of these projects is a citizen-facing AI prototype, codenamed "Humphrey," designed to help users navigate government bureaucracy more efficiently.
The development of Humphrey is currently being spearheaded by the Department for Science, Innovation and Technology (DSIT). Funding for this project comes from the AI Exemplars program. The prototype is expected to be ready for testing within six to twelve months, with a potential nationwide rollout scheduled for the end of 2027, subject to evaluation.
However, the UK government is taking a cautious approach to the deployment of citizen-facing AI. Recognising the "hallucination" problem inherent in large language model (LLM) technology, the government has decided to make Humphrey optional for users. This approach aims to limit user exposure to potential errors early in the deployment of the AI.
To further ensure the safe and reliable use of AI, the UK government has appointed Jade Leung as the Prime Minister's AI Adviser. Leung, an AI governance expert, brings a wealth of experience to the role, having spent over two-and-a-half years at OpenAI, first as governance and policy adviser, then as governance lead. Prior to her time at OpenAI, Leung was head of research and partnership at the Centre for Governance of Artificial Intelligence (GovAI).
In addition to her role at the AI Security Institute, Leung will split her time between advising the Prime Minister on AI and her full-time role at the Institute. Leung's appointment is part of the government's efforts to involve AI governance experts in crafting responsible AI strategies.
Professional sectors using AI, such as legal services, must adhere to strict monitoring, verification, and disclosure policies to catch and mitigate potential hallucinations. Regulatory bodies like the Solicitors Regulation Authority require lawyers using AI tools to monitor for bias and hallucinations, maintaining full professional responsibility for AI-generated content. This ensures compliance with existing laws and ethical codes, indirectly addressing hallucination-related risks by enforcing verification and accountability standards.
Despite the lack of a standalone AI law, the UK government is leveraging existing legal and regulatory frameworks to manage AI risks, including hallucinations. The government is also encouraging collaboration with the world's brightest AI developers to test the value of the latest AI technology.
The AI Exemplar efforts are expected to unlock £45 billion in productivity gains, making the UK a global leader in AI innovation. With the appointment of Jade Leung as the AI Adviser to the Prime Minister, the UK government is committed to addressing AI risks comprehensively while harnessing the power of AI to improve public services and create a more efficient government.
