Government publishes first guidance on AI in social care
The Department for Education and the Department of Health and Social Care have published AI in social care, the first joint guidance on using artificial intelligence (AI) safely in children's and adult social care in England. It is non-statutory and written mainly for local authorities. The government says its principles also apply to other organisations providing social care, including homecare providers.
The guidance sets out four principles.
- Practitioners stay responsible for their records and decisions, and should treat AI output as a first draft.
- AI should not make decisions that directly affect people, including decisions on care planning.
- AI should support better outcomes and free up time for direct care.
- Nobody should enter personal information into public AI tools.
The guidance also says people have the right to refuse AI in their care, and providers should offer a non-AI alternative. It includes governance and practitioner checklists that providers can adapt, and a quick guide for practitioners.
The guidance doesn't change CQC's approach. CQC's statement on AI makes clear that it doesn't approve AI tools, but it expects providers to use AI with human oversight, transparency and strong governance. If you use AI, or are thinking about it, check that you have an AI policy, a data protection impact assessment for each tool, and real human review of anything AI produces.
The guidance applies to England only. Providers in Scotland can refer to the Scottish Government's interim guidance on AI in health and social care, and providers in Wales to Social Care Wales's AI guide for social care.