Leadership in Practice
The UK’s fastest-growing news and insight platform for the care sector.
Practical advice, strategies and insights to help care sector leaders attract, recruit, retain and develop exceptional teams. Our guidance covers everything from writing job adverts that attract the right people and identifying interview red flags to understanding the cost of a toxic team member, with a range of resources to help you build stronger, more effective workforces.
Artificial intelligence is moving rapidly into health and social care. It may help organisations complete administrative work, analyse information, produce draft documents and reduce the time employees spend entering the same information into several systems.
For leaders, the appeal is clear. Care services are under pressure, managers carry significant administrative workloads and frontline teams regularly say they want more time with the people they support.
For workers, the arrival of AI may feel less reassuring. They may wonder whether the organisation intends to reduce jobs, monitor performance or replace professional judgement with an automated system. Some may fear being blamed when technology gets something wrong, while others will worry about confidential information being entered into tools they do not understand.
Telling employees that AI is exciting will not remove these concerns. Leaders need to explain why the technology is being introduced, what it will and will not do and how people will remain involved in every important decision.
A weak introduction begins with a demonstration of what an AI system can do. A stronger introduction begins by explaining the problem the organisation is trying to solve.
Perhaps care workers are spending too much time duplicating records. Managers may be losing hours preparing routine communications, or important information may be difficult to find across several systems. Employees need to understand why AI has been chosen and what evidence supports that decision. The purpose should be specific enough for the team to judge whether the technology has actually improved anything.
“We want to explore whether this can reduce the time you spend producing repetitive summaries” is clearer than, “We are becoming an AI-enabled organisation.”
AI should not be introduced because competitors are using it, because a salesperson gave an impressive demonstration or because leaders are frightened of appearing behind. It should address a real need without creating disproportionate new risks.
If leaders avoid the employment question, workers will answer it themselves. Rumours may spread that administrative roles will disappear, hours will be reduced or employees who struggle with technology will be managed out. Even a small pilot can feel threatening if nobody explains the organisation’s longer-term intentions.
Be as honest as possible. If the purpose is to support the existing workforce rather than reduce it, say so clearly. If particular responsibilities may change, explain what is known and what remains undecided.
Do not promise that AI will never affect any role if you cannot guarantee it. False reassurance may calm the initial meeting, but it will damage trust if the organisation later changes direction. Workers can cope with change more effectively when they receive accurate information, understand the timescale and have a genuine opportunity to influence what happens next.
Teams frequently hear impressive claims about what artificial intelligence can produce. They hear far less about its limitations. AI can generate information that sounds confident but is inaccurate. It can misunderstand context, reproduce bias and omit important details. Its output may be persuasive precisely because it is written fluently.
Care workers must understand that a professional-looking answer is not necessarily a correct one. The technology does not know the person receiving care in the way their regular support team does. It cannot accept responsibility for a decision, notice every subtle change in someone’s behaviour or understand the full consequences of an error.
Human judgement, observation and accountability remain essential. If the organisation cannot explain where human oversight sits, the system is not ready to be used in care.
Care organisations hold highly sensitive information. Employees may be tempted to use publicly available AI tools to summarise notes, draft reports or improve writing without understanding where that information goes.
A vague instruction to “be careful with data” is not enough. Workers need explicit guidance about which tools are approved, what information can be entered and which uses are prohibited. They must understand that names are not the only way somebody can be identified. A combination of details about age, location, diagnosis, circumstances or a particular incident may reveal who the information concerns.
Staff should never be expected to decide individually whether an unapproved AI platform is secure enough for care information.
The Information Commissioner’s Office has warned that a lack of transparency about how organisations use personal information can undermine trust in AI. Providers therefore need proper information governance, appropriate risk assessments and clear explanations for people receiving care as well as employees.
If leaders cannot confidently explain how information is protected, they should pause before implementation.
AI systems are often selected by senior leaders and introduced by people who will never use them during a pressured shift. Frontline workers can identify practical problems that are invisible in a boardroom. They know when information is collected, which tasks cause duplication and where a new process could create additional risk.
Include workers from different roles, shifts and confidence levels. A system tested only by managers who enjoy technology will not reveal how it feels to someone using it alone at night or between tightly scheduled homecare visits.
People receiving care and their families may also need to be involved, particularly where AI affects records, communication, monitoring or decisions about support. Consultation should happen before the final decision, not after the contract has been signed.
Every team contains people who adopt new technology quickly and others who need more time. Both groups provide valuable information. Enthusiastic workers can explore what the system makes possible. Less confident workers are more likely to identify confusing instructions, accessibility problems and areas where the training is insufficient.
If a pilot succeeds only because it was completed by the most technologically capable employees, it has not demonstrated that the system will work across the service. Include different ages, roles, shifts, communication needs and levels of digital experience. Make it safe for participants to say they do not understand something.
A worker who struggles with a poorly designed system is not necessarily resistant to change. The system may be difficult to use.
Introducing AI through an email and a link to an online module is unlikely to build confidence. Workers need protected time to see how the tool operates, practise using it and ask questions. Training should use realistic care scenarios rather than generic examples that bear no resemblance to their work.
Explain common errors and demonstrate how staff should check the output. Show them how to report a problem and what to do when the system produces something unsafe, inappropriate or plainly wrong. Training should also cover when not to use AI.
Employees need to know which tasks require direct professional judgement and where the approved boundaries sit. Confidence is not simply knowing which button to press; it is understanding when the tool should not be trusted.
One of the greatest sources of anxiety is uncertainty about who is responsible when AI contributes to an error. If a system drafts a care note, who checks it? If it summarises information incorrectly, who identifies the mistake? If the recommended wording creates a safeguarding risk, who has authority to reject it?
The answer cannot be that staff should use the technology while leadership remains protected from the consequences. Policies should identify who reviews outputs, who approves final decisions and how errors are recorded and investigated. Workers should never assume that using an approved system transfers responsibility away from the organisation or removes the need for professional judgement.
At the same time, employees should not be punished for identifying a genuine problem with a tool. Early reporting helps an organisation act before the same issue affects more people.
The possibility of reducing time spent writing notes will be attractive to many care teams. However, records are not an administrative inconvenience with no wider purpose.
Care notes support continuity, safeguarding, clinical decisions, communication and accountability. They may later be examined during complaints, investigations, inspections or legal proceedings.
If AI is used to help draft or summarise records, the worker must verify that the final entry is accurate, specific and genuinely reflects what happened. The system should not introduce details, remove uncertainty or turn a worker’s observation into a clinical conclusion they are not qualified to make. Generic, polished records may look impressive while losing the individual detail that makes them useful.
AI should help workers record care more effectively. It should not make every person’s day sound exactly the same.
Workers may worry that new technology will be used to monitor how quickly they complete tasks, where they travel, what they say or how their performance compares with colleagues. If monitoring is part of the system, employers must be transparent about it. Staff should understand what information is collected, why it is necessary, who can see it and how it may be used.
Technology introduced publicly as a way to support care should not quietly become a tool for disciplinary surveillance.
Even where monitoring has a legitimate purpose, excessive data collection can damage trust and encourage defensive behaviour. Employees may focus on satisfying the system rather than responding naturally to the person in front of them. What can be measured is not always what matters most in care.
AI is frequently promoted as a way to save time. Workers will reasonably ask what happens to the time saved.
If every reduction in administration is immediately converted into additional visits, greater workloads or fewer staff, the team will not experience the technology as supportive. They will experience it as another way of extracting more work. Leaders should define the intended benefit. Is the time being returned to direct care? Will handovers become less rushed? Can managers provide better supervision? Will workers finish records within paid hours rather than completing them at home?
The promise that AI will “free staff to care” needs to become visible in daily working conditions. Otherwise, employees may conclude that efficiency applies only to the organisation while pressure continues to increase for them.
AI is a broad term covering many different tools and risks. Trying to transform the entire organisation at once creates confusion and makes it difficult to identify what is working. Begin with a limited, lower-risk use where the purpose and boundaries are clear. Agree what success looks like and monitor the effect on staff time, care quality, accuracy and confidence.
Record problems as carefully as benefits. A tool that saves ten minutes but creates fifteen minutes of checking has not improved efficiency. A system employees avoid or work around is not succeeding because the implementation report says it is live. Only expand after the organisation understands the consequences of the first use.
Care providers are accustomed to assessing risk before changing practice. AI should not be treated differently because it is fashionable.
Workers need to know where to go if the system makes a mistake, produces biased output or appears to compromise someone’s privacy. A general IT helpdesk may not be enough when the concern affects care, safeguarding or professional accountability.
Create a clear reporting route and identify who will review concerns. Make it explicit that staff can stop using the system in a particular situation if they believe it may cause immediate harm.
Leaders should report back on what has changed as a result of employee feedback. If concerns disappear into a form and nobody hears anything further, workers will stop raising them. Psychological safety matters in digital change. People must be able to question the technology without being labelled negative or old-fashioned.
Some highly experienced care workers may feel embarrassed by their lack of digital confidence. They may have spent years developing excellent relationships and sound professional judgement, only to feel suddenly less capable because a new system has arrived.
These employees should not be treated as obstacles. Offer different forms of learning, repeat training where needed and avoid public comparisons between staff. Pairing workers can help, provided it does not turn one employee into permanent unpaid technical support.
Leaders should also recognise accessibility needs. Visual impairment, dyslexia, language differences, disability and limited access to devices can all affect how workers experience a system. An organisation has not modernised successfully if it gains a tool but loses the confidence of experienced people.
The first weeks of implementation will not reveal every effect. Workers may initially follow the process closely, then develop shortcuts as workloads increase. Errors may emerge only in unusual situations, and people receiving care may take time to express how the change feels.
Review the system regularly. Ask whether it saves time, whether staff trust it and whether it has affected relationships or the quality of records. Look for unintended consequences as well as technical failures. Be prepared to change the process, provide further training or stop using the tool if the risks outweigh the benefits.
Continuing with unsuitable technology because the organisation has already paid for it is not innovation. It is refusing to learn.
Care workers are not necessarily frightened of technology. Many already use digital systems every day and will welcome anything that removes unnecessary duplication or gives them more time with people.
What frightens teams is uncertainty. They become anxious when leaders cannot explain whether jobs are safe, how data will be protected, who is responsible for errors or whether “saving time” really means increasing workloads.
Trust grows when organisations answer those questions honestly and involve workers before decisions are finalised.
AI has the potential to support care, but it should remain a tool used by people—not a mysterious authority imposed upon them. Introduce it carefully, explain it clearly and keep human judgement where it belongs: at the centre of every decision affecting someone’s life.
Join organisations across health and social care that are investing in happier, healthier workplaces with the Daily Round Wellbeing Employer programme.
Download our Employer Wellbeing Brochure to discover what’s included, or become a Daily Round Wellbeing Employer today.
Posted by:
K Jadon
Editorial Assistant – The Daily Round
Busy day? We’ve got you covered. Get FREE daily news, practical insights, opportunities and wellbeing content, carefully curated for busy health and social care professionals. Just enter your details below to get The Daily Briefing delivered straight to your inbox.
newsdesk@dailyround.news