Evaluating the current state of AI throughout UK local government
Local government is being firmly encouraged to accelerate the adoption of artificial intelligence (AI) in a bid to improve efficiency and productivity. While the potential is compelling, not least the chance to empower existing staff with the tools and insight required to meet escalating service demands, perspective is key.
Following technology exploration, with significant prototyping and ideation, as impactful implementation begins to gather pace throughout local government, tried and trusted processes for both service delivery and application development cannot be overlooked.
AI will deliver innovation at a new pace, but it is important to step back from the hype and accurately assess what can safely be achieved today. Rick Hassard, Director of Engineering and member of the AI Board at Idox explains the importance of local authority-specific AI development and close collaboration between suppliers and local authority experts, especially within regulated areas such as planning.
Increasing pressure to adopt AI
AI is on every local authority agenda. With a projected ยฃ8bn funding black hole by 2028/29 and the threat of Section 114 notices adding further strain, there is no doubt that change is required. Every aspect of local government spending is facing review, and with service demands increasing, especially in areas such as planning and social care, local authorities are experiencing enormous pressure to deliver more with less.
The shortage of skilled staff is adding to the challenge. While skill shortages vary depending on each council and region, the LGAโs workforce survey suggests recruitment difficulties in occupations such as planning, legal, digital, environmental health, finance, and childrenโs and adultโs social workers. The issue is particularly pressing within planning, with around eight out of 10 planning departments short of staff.
With only 20 per cent of departments having enough planners to handle applications, it is inevitable that only one in five applications for major projects were decided within the 13-week statutory period in the past three months. The government has pledged to recruit 300 council planners, but the study points out that this only plugs 15 per cent of the shortfall.
These issues would be challenging within a stable environment. For local authorities facing an increase in total expenditure on adult social care, government targets to build 1.5 million new homes and a shift towards devolution, where is the time or resources required to achieve digital transformation or embrace innovation such as AI?
AIโs potential to improve efficiency and productivity
The government is firmly committed to the use of technology to improve performance, efficiency and productivity across both central and local government. Every week seemingly brings a new announcement of AI adoption in an array of areas, with pothole management, traffic management, benchmarking attendance in schools and minute taking in meetings just a few examples. The objective in each case is to improve the performance and efficiency of existing staff by streamlining processes and removing mundane tasks.
One of the biggest areas of focus is, of course, the planning sector. Recently, the government announced the development of the Extract AI assistant to support faster planning decisions. In test trials across Hillingdon, Nuneaton & Bedworth, and Exeter councils, Extract digitised planning records, including maps, in just three minutes each โ compared to the 1โ2 hours it typically takes manually. This means Extract could process around 100 planning records a day โ significantly speeding up the process.
The potential is clear; local government experts using the latest iterations of GenAI tools, including LLMs, achieve better quality, speed and reduced costs. The Tony Blair Institute worked with a local council in the UK and estimated that AI could be applied to 26 per cent of tasks, resulting in a saving of one million work hours, or ยฃ30m in financial costs, each year. Scaled up nationally, this would mean around ยฃ8bn in cost savings.
AI could be applied to 26 per cent of tasks, resulting in a saving of one million work hours, or ยฃ30m in financial costs, each year.
Collaborative approach to AI evolution
This is not a one-way process. It is essential that local government is an active participant in the AI evolution to ensure products are developed that not only meet operational needs but do so safely and effectively. Local authority experts understand the regulation and how systems are used within that regulatory framework. Their input is key to ensuring AI is directed towards the areas that add quantifiable value to existing staff.
Environmental Health inspectors, for example, are in short supply, so it makes sense to use AI to support an intelligence-led, risk-based approach to inspections that maximises the value of these skilled experts. Similarly, within planning, there are areas of work that do not require a skilled building control surveyor or planning officer and the use of AI to streamline and automate these processes can release skilled experts to focus on more complex, valuable activity.
Close collaboration between software developers and local authority experts will accelerate the development of innovations that can be deployed at scale. In addition to confirming the priority areas for development, human oversight from local authority experts, sanity checking the veracity of AI responses, for example, are key to building both accuracy in the model and confidence in the technology. Furthermore, working with existing technology partners with an extensive understanding of and commitment to the UK local government environment will also safeguard valuable data resources.
Current state of AI evolution
But where do these pilot and scoping developments leave local authorities? There is a gap between a concept or pilot and widescale operational use. Are these solutions available for deployment? Can they be used alongside existing systems? What integration is required?
How much will it cost? Going from proof of concept to production requires investment โ and how is that investment to be funded? Where is the business case, the proven value, even the trusted deployment partners?
While consumer GenAI tools are being subsidised by the tech industry, the reality is that AI can be costly to develop and implement. In addition, rising concerns about hallucinations, bias and fairness can undermine confidence both amongst end users and, potentially, citizens who will be worried about the implications for service delivery. This is of particular importance in areas driven by regulation, including planning and public protection. From assessing planning applications to managing health inspections, it is vital to be 100% confident in the accuracy and relevance of information provided by AI before any widescale deployment can be achieved.
Local government needs a voice. Local authorities need to understand where and how central government-driven projects with tech suppliers fit into digital transformation strategies. They need clarity about the risks and rewards associated with data ownership and usage, and the difference between generic AI tools and those designed specifically for UK local government. This is where existing suppliers have an essential role to play in reassuring local government that AI innovations are being built to meet specific, UK relevant requirements.
A voice for local government
Idoxโs expertise and years of experience provide a level of insight and understanding into the specific needs of UK local authorities, ensuring solutions are shaped by real operational requirements and grounded in the day-to-day realities of service delivery. We know where innovative technologies fit into existing systems. Technologies such as Retrieval-augmented generation (RAG), which allows LLMs to retrieve and incorporate new information, and help prevent hallucinations by grounding information in trusted data sources, such as planning specific information, to elevate relevance and accuracy.
Or sentiment analysis of local authority data โ whereby the meaning embedded within the text is assessed to gain greater insight into emotional context โ which can be used to gain customer feedback and improve customer service. Within planning, for example, sentiment analysis of resident responses to planning application statements can provide immediate insight into whether the perception is negative or positive.
At the heart of successful AI adoption is ensuring local government has access to the right AI models that reflect the specific needs of the UK government. Rather than generic LLMs or those trained on government data from Australia, New Zealand, Canada or the US, LLMs specifically trained on UK government data and developed according to UK regulatory requirements are key to delivering the accuracy and certainty required, especially within regulatory areas.
Conclusion
The excitement surrounding AI is not misplaced. There are undoubted opportunities to address many of the challenges currently facing local government, not least supporting existing staff to deliver more with less. However, despite the extraordinary pace of change, there is no need to feel rushed into investment.
Indeed, the public sector has the opportunity to learn from the often painful lessons learnt from private sector companies that have rushed headlong into AI development without the essential checks and balances. Aligning development with clearly understood goals is the priority, but companies that have relied on inaccurate data and failed to involve humans in the development process have paid a hefty price.
Idox is actively at work in AI and continues to seek collaboration with customers. We recognise the significant opportunity to support local government in addressing workforce and budgetary pressures. Our research into humanโAI collaboration focuses on streamlining routine tasks, enabling skilled professionals to concentrate on more complex and high-value challenges. Early-stage proof of concepts from our R&D efforts are already demonstrating strong potential, and we are now working closely with customers to bring these innovations into production.