Social care providers need more support to help them implement artificial intelligence (AI) safely, a charity has urged.

In a new report from Care England, titled ‘AI has arrived in social care: supporting providers with adoption’, the charity highlights that there is a ‘major challenge’ faced by care leaders where AI governance is concerned.

A survey of 26 care leaders found that most use AI daily and feel somewhat or very confident with the technology despite low levels of formal training.

But respondents ‘consistently identified the lack of clarity and misalignment in understanding around AI governance requirements as the greatest challenge to successful AI adoption’.

The survey revealed that many have adopted their own policies, frameworks and internal policies as a result. However, they ‘do not feel empowered by regulators and commissioners’ in this area.

According to the research from the Institute for Ethics in AI at the University of Oxford, Serene Care, the Digital Care Hub and Care England, respondents expressed the need for more regulatory clarity, improved training opportunities and peer support.

Furthermore, it found that new responsibilities such as a greater administrative burden have arisen due to the use of AI. For example, care leaders are required to check AI-generated information for accuracy, bias or discrimination and completeness.

Concerns were voiced about data privacy, the impact on person-centred care and potential overreliance on AI. The survey also identified time and affordability as obstacles to AI adoption.

As part of its recommendations, the report identifies the need for a ‘more supportive, transparent and collaborative approach’ between local commissioners and regulators and care leaders.

Additionally, it calls for more funding for AI and digital training, as well as urging that the Casey Commission considers the role of AI in the future of social care in England.

‘The future needs with demographic aging will be significant, we need to find ways to deliver better outcomes and support care providers, and this has to happen in a supportive, human-centric way, that does not only focus on processes in care services, but also the outcomes for the people using these services and those working in them’, it reads.