How AI is used in social work

As a profession, social care is entering into a necessarily slow but promising embrace with AI.

Slow because social care is, Tommy emphasises, very different from healthcare. Visiting the doctor or a hospital is often about seeking a diagnosis and, hopefully, a cure.

Social care, through understanding a person’s unique life story and circumstances, aims to provide support so that the person can live a better life. ‘There isn’t a five-aside football injury to be fixed’, he says. ‘There isn’t an operation to be had or a course of antibiotics. If you're talking about someone who has autism, for example, there's no pill or injection that changes that. That is them. It's interwoven into their personality and their outlook on life.

Defining social work and social care

‘Social work is hard to pin down with a single definition’, Tommy explains. It’s not just about helping with tasks like washing or dressing. At its core, it is a non-medical service that supports people whose health or disabilities affect their daily lives. Often delivered informally by families and carers, social work is really about creating the conditions for people to live fully and independently. It helps ensure everyone can take part in society, no matter their circumstances.’

This means, Tommy suggests, that how AI might ultimately be used in social care may be quite different from how it's deployed in healthcare. In healthcare, there’s often a population-scale body of evidence, data and research about diseases, illness and potential cures. And, as we all know, AI is great at spotting patterns in vast amounts of data. Social care is, however, much more personal.

‘I’m not saying one is better than the other’, Tommy says. ‘But healthcare is quantitative. We have “do this and do that”…We've got a track record, history. There are permutations, but ultimately we can get from point A to point B, and we know what to do.

‘Social care is qualitative…A person's life and their story, triggers, interests and nuances. Yours are completely different to mine, as are everyone else’s in the entire world. So, you have to look at technology through that lens.’

Why don’t you tell us about your career?

I started out in youth careers advice, then moved into children’s social care before retraining with a master’s in social work. I have been a registered social worker for over 15 years, working across a range of services from early intervention to complex adult safeguarding. 
I spent several years in London supporting hospital discharge planning, helping people move safely between the NHS and social care systems. I now lead on social care for BCS and the Digitising Social Care programme, which brings together the Department of Health and Social Care and NHS England. People come to social care for many reasons — ageing, disability, mental health, or needing help with daily life. My work focuses on ensuring they receive the right support at the right time in the right way.

How did you make the shift from being a social worker to digital?
After years in frontline social work, I felt the emotional toll and saw how poor digital systems made things harder. I wanted a job that meant something, one that aligned with my values and my interest in people and their stories... supporting others to live the lives they want. This felt powerful and worthwhile. So, in 2017, I joined an NHS Digital fellowship, shifting into clinical informatics.

It was a steep learning curve, but I saw the value of bringing social care expertise into digital roles. That realisation led me to work with care providers, then into national leadership — connecting practice with policy, data and technology to improve social care’s digital maturity. It is work I feel proud of, and it continues to challenge and shape me.

Let’s jump forward in time and imagine what it might feel like to work in a social care system that has been digitally transformed.

If we imagine the future of social care, technology should not replace human connection, but enhance it. It should amplify the voices, needs and experiences of people drawing on care and support. Rather than treating them as data points, it should tell their story in a way that professionals can access instantly and use meaningfully. This is not about medicalising people, but about recognising their individuality. True person-centred technology captures what matters most: how someone wants to live, what support they need, and when and how they want it. Whether services are state-funded or self-funded, the goal is the same; technology that is accountable to the people it serves, enabling them to live fully and with dignity. It is their story, and tech should help tell it.

Talk to us about digital maturity in social care. Is it being held back? Where is it relative to other sectors?

That’s a good question. The thing is, social care is massive. It employs around 1.6 million people, and it is not one single system. You have local authorities, private care providers, individuals and families all playing different roles. So, when we talk about digital maturity, the answer really depends on where you look.

Some organisations are ahead of the curve, digitally agile and thinking smart about connected care. In fact, some even find the NHS hard to work with because it is behind in some areas, which flips the usual stereotype. But on the other hand, a lot of care providers have only just moved from paper to digital records in the last few years.

Broadly speaking, social care is probably five or six years behind the NHS in digital terms. That said, there is a silver lining. Social care does not have the same legacy tech issues the NHS does. It is more of a greenfield site, which means there is room to build better from the ground up.

Additionally, because many social care organisations are independent they are not tied to rigid national frameworks. That gives them more flexibility. So, while the sector is behind now I would not be surprised if, in 10 years, it overtakes the NHS in some areas.

How easy is it to move data between organisations?

It’s not easy to move data around in social care. The sector is incredibly varied, with large local authorities, small care providers and unpaid carers all working differently. There is no single system or standard. Some organisations are digitally advanced, but many are not and lack dedicated digital teams. That means they often do not know what they need or where to go for support.

Work is underway, like the Federated Data Platform, to build the foundations for better data sharing. But it will take years to fully implement. Unlike banking, which had a clear societal push for digitisation, social care has not had that same unified momentum. While the potential is there, moving data around remains a major challenge.

Talk to us about who should be involved in designing AI projects and products.

Designing AI for social workers should involve a mix of people. That includes the workforce across different roles, regions and settings, as well as people who draw on care and their families. It sounds simple, but it isn’t. You have to listen carefully to what people actually want to use, not just what sounds clever or new.

For you

Be part of something bigger, join BCS, The Chartered Institute for IT.

There is often a rush to build something shiny, but sometimes what is being built already exists. I remember visiting a supported living setting in Rotherham where we showed residents four types of care tech. One was a robot that let you call family. A resident in their seventies looked at it and said, ‘Why are you doing this? I have a mobile phone that does the job.’ That stuck with me.

So, if you are building AI for social care, it has to be grounded in proper user research. Ideally, there should be evidence that people were involved in shaping it. Without that, you risk creating something that adds complexity instead of value. In fact, it would be helpful to see standards like those from ISO requiring developers to demonstrate that they have engaged with users meaningfully. That kind of accountability could stop a lot of poorly designed products from ever reaching the market.

Where is AI making a difference today — right now?

That depends on the type of AI you are talking about. Predictive analytics have been around for a while. Local authorities use algorithms and large data sets to help plan services and allocate budgets based on population health. That kind of AI is already quite widespread.

Then there is generative AI. You will find care workers, managers and local authority staff using tools like ChatGPT in different ways — sometimes to draft care plans or write notes. There are risks with that, especially when people don’t realise that they should not be putting personal information into those tools.

We are also seeing more use of tools like Co-pilot for things like supervision notes and meeting summaries, just like you and I might use them. Then there are more tailored tools, like Magic Notes, which is designed for social workers and occupational therapists. It uses voice-to-text during assessments and creates an action plan on the spot.

Another interesting example is PainChek. It uses facial recognition to detect pain in people who cannot express it themselves. It learns their facial movements and helps identify when they might be in pain. It’s already being used in care settings across the country.
Summing up social care as a profession, Tommy says: ‘We're not here to fix people. We're here to support people to live a gloriously ordinary life — just like you and I.’

For emphasis, he repeats himself: ‘Gloriously ordinary.’