As interest in AI continues to grow, questions are being asked about its appropriate role in adult social care. With longstanding workforce pressures and increasing demand, AI is often viewed as a potential tool to improve efficiency. But its development and deployment must be handled with caution and care.
That is why the Responsible Use of Generative AI in Adult Social Care project – jointly led by the Institute for Ethics in AI at Oxford University, Digital Care Hub and Casson Consulting – set out to build sufficient consensus on what good looks like across a diverse range of stakeholders, including people drawing on and working in social care, technology suppliers, commissioners and policy makers. This work is critical if all stakeholders are to move forward with shared understanding and confidence.
The project acknowledges that generative AI has potential in the sector, particularly in reducing administrative burdens. Tasks such as summarising documents or drafting standard communications are areas where AI could offer support. In freeing up time, it may allow care teams to focus more on delivering person-centred care.
However, the project also highlights significant risks. Current AI tools can produce inaccurate, misleading or biased outputs, especially when applied to complex or sensitive contexts. There are serious concerns about using generative AI to develop care plans or make decisions about people’s needs, given the risk of errors and the importance of professional judgement.
The tension between efficiency and person-led care is a core concern. Poorly implemented AI could erode trust, reduce autonomy for care workers and compromise the dignity and rights of those drawing on care. If AI is to have a place in social care, it must support – not replace – the human relationships at the heart of the sector.
To address these concerns, the AI in social care project issued a series of calls to action. These are intended to help organisations explore AI safely and responsibly, while also encouraging innovation that is grounded in social care values. One of the most pressing priorities is to tackle bias in AI systems. If algorithms are built on unrepresentative data or without input from diverse users, they risk reinforcing existing inequalities. Tackling this requires inclusive design, transparent processes and robust testing across a range of real-world contexts.
Another priority is the need for clearer guidelines and accountability. Care providers need support to assess the appropriateness and risks of AI tools. Technology suppliers need consistent standards to ensure their products are safe, transparent and fit for purpose. Regulation, procurement and commissioning must all play a role in embedding safeguards.
To support this, the project developed a framework for responsible use. It outlines six principles – fairness, transparency, accountability, safety, empowerment and sustainability – alongside desired outcomes. The aim is to provide a practical and ethical foundation for anyone exploring AI in social care.
Reaching a shared understanding is essential. Without it, the sector risks inconsistent and potentially harmful use of AI. But with consensus, it becomes possible to shape technology that enhances, rather than undermines, high-quality care. AI should not be something imposed on the sector. It should be developed with social care – informed by the expertise of care providers, the voices of people with lived experience, and the practical realities of frontline delivery.
AI is not a silver bullet, and it cannot fix the structural challenges facing social care. But if used responsibly, it may become a valuable tool – one that supports sustainability, without compromising on values.
Right now, the pressure on social care is enormous. We have got an ageing population, more complex needs and a workforce that is stretched thin. If AI is used properly, it will take some of the weight off by taking over the boring admin, spotting early signs of health issues, or even just helping keep things consistent behind the scenes. Let’s be clear, AI is not about replacing anyone, it is about making sure care staff have more time to do the real work, talking, listening, connecting. That is the heart of it, augmentation, not automation. But we have got to go carefully. AI that is not tested properly, that ignores bias or that has dropped in without asking people who will use it can do more harm than good. The people delivering and receiving care need to be part of shaping it from day one. So maybe the question is not what AI can do in care, but what does care really need right now, and how AI can humbly serve that, without getting in the way. Getting that balance right is not easy, but with the right conversations, collaboration and a bit of lived experience, it is absolutely possible. I have seen firsthand how thoughtful use of AI can make a real difference, and I am always happy to chat with others exploring these exciting opportunities. Stephen Kenealy, Co-Founder and Chief Commercial Officer, Call The AI. @Stephen-Kenealy-Call-The-AI At its best, AI can provide huge benefits for providers, from freeing up frontline staff from administrative overload, through providing 24/7, high-quality real-time support to those delivering care. Tools designed specifically for the care sector can be game-changing for everyone involved in providing and receiving care services, including providing support with care planning and risk management, improving the quality of care documentation and providing best practice guidance. Effective use of AI gives care teams more time to do what matters most; connect with and care for people. AI-driven insights can also help managers anticipate issues like falls or health deterioration, enabling earlier, preventative interventions. That said, poorly designed, generic AI tools can create blind spots, introduce bias or undermine trust – particularly if they are implemented without involving staff and service users. Data privacy, accountability and consent must be core to every AI initiative. To help rather than hinder, AI must be co-produced with the people who use and deliver care. It should enhance –not replace – human judgement and human touch. It must be transparent, secure and effective. Most importantly, it must support a philosophy of care that is person-centred, dignified and inclusive. If AI is used to cut corners or de-skill the workforce, it will, quite rightly, be resisted. If it is introduced with humility, consultation and a clear focus on improving outcomes, AI can be a powerful ally in creating a more resilient and compassionate care system. Hannah Morgan, Chief Executive Officer, CareBrain. X: @Hannah-Morgan-CareBrain Email: [email protected]
AI could really change the game for social care, but only if it is guided by empathy, not efficiency alone. Technology can be brilliant, but in this sector, there is a real risk it is seen as a shortcut. Let’s be honest, care is about people, about relationships, trust and being present. That cannot be automated. What AI can do, though, is support the amazing carers and nurses already holding the system together.
Email: [email protected]
AI holds enormous potential to transform social care when applied in a way that augments, rather than replaces, the human heart of care. The sector is under immense pressure – rising demand, a stretched workforce and tightening budgets. AI can ease some of that burden by automating repetitive tasks, supporting decision-making and enhancing the quality and consistency of care – but it must be deployed ethically.