North Yorkshire Council
Children and Families Overview and Scrutiny Committee
17th December 2025
Approach to, and use of, AI in Children’s Services
Report of the Corporate Director Children and Young People’s Service
1.0 Purpose of report
1.1 The purpose of this report is to provide an overview of the current approach and use of AI in Children’s Services and how this correlates with the wider council’s approach.
2.0 Summary
2.1 This report details the approach towards the use of AI in Children’s Services and some of the products and functionality being used.
2.2 The report covers detail on different types of AI, Ethical Implications, Benefits and Risk, Workforce Upskilling and Practical examples.
3.0 Background
3.1 Our approach to AI has been cautious in some areas and bold in others.
3.2 We are recognised in the sector as leading many of the developments of AI in Social Work (in its widest sense). Collectively we have won many national awards in the past 2 years for our work including AI Initiative of the Year, iNetwork’s Outstanding Contribution Award and the iStand UK award. We were also invited to the Microsoft Summit in Dublin to share what Microsoft described as our “cutting edge” innovation in reimagining case management.
3.3 Children and Young People’s services work closely with other corporate departments such as transformation, technology and data and insight to explore, test and utilise the opportunities AI brings, whilst being mindful of, and cautious about the risk areas associated with AI such as Ethics, Bias and Information Security.
4.0 What is AI?
4.1 AI is often talked about as being one thing, but in reality, it’s lots of different things. There are well known large language models such as Chat GPT and Co-Pilot which can answer questions based on huge information libraries but also AI Voice to text transcription that can record a meeting and write up meeting minutes and summaries like Magic Notes and Minute AI. In addition to these, there are many other types of AI all commonly referred to under the term AI.
As an organisation we want to harness the opportunities all of these different functionalities offer and also to educate and highlight our workforce that AI isn’t one thing, as well as teaching them how to use it responsibly, safely and ethically.
5.0 Ethical implications
5.1 The Transformation and Innovation Team have spearheaded our approach to ethically informed approaches to AI. Whenever technology is being developed or purchased, we routinely complete Data Protection Impact Assessments, Climate Change Impact Assessments and Equality Impact Assessments. In addition to this, where AI is involved, we also complete AI Ethical Impact Assessments.
The team has developed an AI Ethics impact assessment, aligning closely to international best practice standards which ensures we fully consider and explore implications in the following categories:
• Human Agency, Liberty and Dignity
• Technical Robustness and Safety
• Privacy and Data Governance
• Transparency
• Diversity, Non-Discrimination and Fairness
• Individual, Societal and Environmental Wellbeing
• Accountability
• Resources
All work is robustly tested and checked through each directorate’s reporting mechanisms and corporately to ensure visibility.
In October 2024 we hosted a Social Care AI Discovery day to help the workforce understand the technology, opportunities, risk and to help us shape the direction of developments and innovation. Days like this have been critical in helping us embrace new opportunities with our workforce rather than imposing technology solutions on them.
6.0 Benefits and risk
6.1 There are benefits and risks involved with AI, especially when considering some of the most sensitive information we store.
The benefits can be significant. For example, using AI tools such as Magic Notes, which capture audio recordings, social work visits and meetings, and create good quality minutes and summaries of the conversations was shown to save our social workers up to 1 day per week in admin time. This is time that they can reinvest in working directly with children and their families.
There are also many risks involved in not robustly testing or responsibly using AI. This could be things like not recognising the bias of the models used to train AI, or might be in not robustly testing the flow of information from our staff mobile phones, to Magic Notes’ cloud based processing and then receiving the information back into our North Yorkshire network or the risk of people carelessly putting sensitive information into insecure Large Language models such as Chat GPT.
To mitigate these risks and leverage the significant benefits we have a corporate approach that brings together transformation, data governance, information security and practice which all need to be satisfied before we purchase or develop AI tools for the workforce to use.
7.0 Workforce upskilling
Traditionally, Local Authorities have relied on buying a small number of products. This can lead to difficulties around migration, vendor lock in and capabilities that are limited to the functionality of the individual product.
Given the new landscape as a result of AI developments we are seeing that there are many different “products” available, which cover a wide spectrum of AI functionality.
To avoid some of these legacy issues, our approach to new technologies is to focus on functionality. This focus means we are teaching people how to use the functionality rather than a specific product and we are teaching them to use lots of different things. An example of this is teaching people to ask questions of a large language model to get the best answer when exploring complex information. We are doing this with Co-Pilot and Policy Buddy but the skills you use to create a good prompt can equally help you in using Chat GPT or Gemini.
This approach mirrors how we interact with technology in every day life. Many of us will have 20+ apps on our phones, tablets and TV’s. We have learnt to be able to navigate multiple apps, each performing specific tasks because those apps are useful to us. We don’t have or need one app to manage our social media, online banking and to be a calculator. We need individual ones that are optimised to do their specific task well.
By teaching people to use the functionality rather than the product it also gives us the freedom to switch and change products as and when better/cheaper products come to the market.
8.0 Practical Examples
8.1 Some practical examples of products in CYPS are:
8.2 FinYd – Reimagining Case Management
This is our flagship development here in North Yorkshire. It is a tool that brings information about a child together into one place. Traditionally, staff would have to navigate between different parts of our case management system to find information, which represents lost time and frustration.By bringing this information together into one place, we can search and find the relevant information in seconds.
The tool also enables us to identify key locations on a map which might be important to a child. This could be a distant relative who could help us to keep a child safe or potentially care for a child if they can’t continue to live at home or it might be to help us identify risky addresses associated with a child.
We are able to see, through timelines of involvement the activity that is taking place for a child. An increase in this activity might mean risks are increasing and a decrease in activity might identify that a support plan is working and partner agencies are becoming less worried.
We also use AI to automatically generate ecomaps around a child. The technology reads every word that is written about a child and then creates a network of people which will include family members but also football coaches, dance teachers, neighbours and friends who can help us to keep children safe or could potentially care for a child if they can’t continue to live at home.
This technology is groundbreaking and is already sought after by other local authorities which brings with it the opportunity to share learning and potentially to earn income for North Yorkshire.
An example of an AI generated eco map is below:
8.3 Magic Notes
Magic Notes is a product which captures conversations and writes minutes and summaries of meetings, social work visits, assessment sessions and telephone calls. It enables practitioners to fully focus on the conversation/meeting rather than trying to take notes and enhances their ability to pick up on non-verbal cues such as body language, home conditions etc.
Initial testing of Magic Notes was that it saved significant time and captured more information than if a person did this. We found in testing that users reported saving up to 1 day per week in admin time.
Whilst the product comes with a cost, the return on investment is significant, however, we are also testing other, similar technologies for future use cases. This signifies our long-term commitment to the use cases, not necessarily to the product. It also places the emphasis on suppliers to continue to build on their products to stay ahead of their competition.
8.4 Policy Buddy
Policy Buddy is a bespoke large language model social work assistant. It looks and feels like Chat GPT but rather than looking at the internet for its answers it only looks at a document library that we have created. In the library is all legislation relating to children, statutory guidance and our policies and procedures.
What this means in practice is that in seconds, a practitioner can ask a question, and it will give an accurate and clear answer.
Our user testing of Policy Buddy was so successful that we were inundated with requests from our staff to have access to it. This was a sign that we had created something that had a very specific use case, but which was very effective at meeting that need.
As we rolled it out to the whole workforce, we then saw additional benefits such as our neurodiverse staff telling us it was a “neurodiverse social workers bestie” because they can control how it provides its answers.
We also saw people using it creatively to create superhero guides to help children understand the child protection process and translating information and answers into hundreds of different languages in seconds.
Policy Buddy has become so popular that the workforce have nicknamed it Polly. Below is an example of the start of a superhero guide aimed at an 8-year-old boy to help him understand the looked after review process. This was generated in seconds, derived from legislation and our policy and procedure.
 |
9.0 Contribution to Council priorities
9.1 This work contributes towards the council priority Safe, Healthy and Living Well predominantly but also contributes to One Council through multi directorate collaboration.
10.0 Alternative options considered
10.1 Whilst this briefing paper is for information and consideration only, options appraisals take place for all innovation, technology and AI work
11.0 Impact on other services/organisations
11.1 Due to the multi directorate collaboration required for all of the work in this paper, there are positive impacts on other services within North Yorkshire. We share all of our learning and benefit from other directorates sharing their learning and knowledge with us.
12.0 Financial implications
12.1 Not applicable.
13.0 Legal implications
13.1 Not applicable.
14.0 Equalities implications
14.1 This report highlights that equalities implications are considered for all innovation work.
15.0 Climate change implications
15.1 This report highlights that climate change implications are considered for all innovation work.
16.0 ICT implications
16.1 This report highlights that multi directorate collaboration is required to achieve this work. This includes ICT.
17.0 Reasons for recommendations
17.1 This report is brought to the committee at their request to highlight how officers within Children’s Services are utilising AI.
|
18.0 |
Recommendation
|
|
|
i) The committee is asked to note the report. |
|
|
|
Pete Thorpe
Corporate Director Children and Young People’s Service
County Hall
Northallerton
3 December 2025
Report Authors – Jonny Hoyle (Children and Young People’s Service) and Cath Ritchie (Transformation)
Presenter of Report – Jonny Hoyle – Group Manager, Leaving Care, Refugees and Asylum Seeking Children
Note: Members are invited to contact the author in advance of the meeting with any detailed queries or questions.