Government issues AI guidance for adult social care providers

Published by the Department of Health and Social Care (DHSC) on 29 September, the guidance provides an introduction to AI in adult social care and sets out best practice for care providers.

The guidance notes that although many adult social care providers remain in the “early stages of exploring AI”, a range of solutions are already being used to support the direct delivery of care and support, including:

• sensor-based technologies, including acoustic monitoring 
• chatbots 
• facial recognition technology 
• data collection and analytics technologies

Generative AI is also being used to create individual care plans and care assessments.

The guidance states: “AI tools can fast-track high-workload tasks such as auditing and writing care plans, daily monitoring and logging data. These AI-backed care planning software technologies can also improve understanding and communication within care planning documentation.”

However, the guidance warns that evidence that generative AI can create “truly personalised care plans” is currently limited, adding that where AI is used in care planning, a member of care staff must review the plan to ensure it is accurate.

The Government notes that organisations should include human review of AI-generated outputs to ensure accuracy and fairness.

Providers should also consider having an AI policy, “even if your policy is that staff cannot use AI for work purposes”. Such a policy could set out how different roles should interact with AI and include a review checklist to help plan continual human review.

The guidance highlights a number of other areas for providers to consider, including data protection and the quality of data supplied to AI systems.

It says organisations using AI tools must comply with UK General Data Protection Regulation (GDPR) requirements and should consider where data is being processed and how people will be told about the use of their data.

The guidance adds: “If you’re using a free or online tool, you should not use personally identifiable information because you would be providing confidential information to a third party. As you would not have a contract with the technology supplier, you will have no way of guaranteeing the safety of that information or how the data will be used.”

The guidance also advises providers to be aware of and mitigate bias. It notes that AI can show biases based on the data on which it is trained, potentially reinforcing stereotypes, causing discrimination and affecting decisions based on biased outputs.

Finally, the guidance notes that organisations should continuously improve their use of AI systems by seeking feedback from human reviewers and, where possible, the AI system itself. As AI evolves and organisational needs change, providers should adapt the review process accordingly.

Lottie Winson