Nottinghamshire County Council set out its current approach to the use of artificial intelligence (AI), outlining how it is being used in frontline services, the governance arrangements being developed and safeguards being put in place to manage potential risks.
The report, which preceded the March 12 summary committee meeting, explained that AI is not a single technology, but rather a collection of tools designed to perform tasks that typically require human intelligence. These include machine learning, which allows systems to identify patterns in data to support predictions and decision-making, and generative AI, which allows systems to create or summarize content in response to prompts.
The council said that while AI offers opportunities to improve services and productivity, it also comes with legal, ethical, operational and security risks. As a local authority with large amounts of sensitive data and responsibility for statutory services such as adult social care, children’s services, education and planning, you need to ensure compliance with the UK GDPR and maintain a strong governance structure.
The report highlights concerns including risks such as algorithmic bias, inaccurate or fabricated content generated by AI systems, over-reliance on automated output by staff, and increased exposure to cyber threats. It also mentions the need to prevent uncontrolled use of AI tools and ensure that projects are driven by service needs rather than technology alone.
An AI policy has been drafted and is currently going through the council’s internal approval process. Governance of AI initiatives has traditionally been overseen through digital boards and program boards, but we plan to introduce a three-tier model consisting of an AI governance board for strategic oversight and ethics, an AI steering group to align activities with business priorities, and an AI center of excellence to provide technical assurance and standards. The council also participates in national AI forums, including the Local Government Association AI Group, and has developed internal training for staff on the use of tools such as Microsoft Copilot.
• Pothole protest at West Bridgford courthouse.
One of the most developed use cases is adult social care and health. In early 2025, the City Council began piloting an AI-assisted transcription tool known as Magic Notes. After review, the tool was approved for widespread deployment in September 2025. Magic Notes captures conversations during ratings and reviews, creates verbatim transcripts, and generates structured summaries. Practitioners will review it before uploading it to the council’s Mosaic system.
The stated aim is to reduce administrative burden, improve the quality and timeliness of case records, and allow staff to spend more time supporting residents and developing care packages.
Comparing the pre- and post-implementation surveys, we find that the percentage of staff spending more than half of their time on administration has decreased from 56 percent to 23 percent.
The percentage of staff who spent more than an hour writing notes per visit decreased from 59 percent to 28 percent, while the number of staff who submitted notes within one day increased from 34 percent to 73 percent.
The percentage of staff who rated the quality of their conversations with the people they support as 8 out of 10 or higher increased from 35 percent to 72 percent.
In parallel, the city council ran a 2025 pilot of Microsoft Copilot involving 300 users. Within Adult Social Care and Health, we analyzed 182 usage entries and recorded an estimated total time savings of 378.3 hours, with a median time saved per entry of 1.33 hours and an average efficiency increase of 72%. Meetings accounted for the largest cumulative time savings, primarily through automated note-taking and action summarization, but applications and automated work also provided significant benefits.
Looking ahead, the council says it will continue to take a cautious and ambitious approach, focusing on proven technologies that are supported by business cases and benefit tracking, rather than adopting untested innovations. Work is underway with Access Group to explore how AI can support Mosaic and related products in areas such as accelerating care package reviews and predictive reporting. Consideration is also being given to AI capabilities within the SAP system and a digital EHCP assistant developed by Somerset County Council.
The report outlines potential future applications in line with the council’s strategic priorities, including multi-agency data analysis to early identify families in need of support, AI-assisted scheduling and care planning, digital tools for SEND support, community insight dashboards, and systems to support planning and infrastructure forecasting. However, these are not confirmed projects and are presented as objects of investigation.
There are no direct financial or legal implications arising from the report itself as it is for the purposes of the Overview Committee. Future AI implementations not yet reflected in Congressional budgets will require separate approval, including details of revenue and capital costs. The report confirms that procured AI technology must comply with the Procurement Act 2023, the Procurement Regulations 2024 and the UK’s GDPR requirements.
Overview Committee members are asked to consider and comment on the Council’s current approach. The report notes that AI is expected to assist authorities in their efforts to implement local government restructuring.
