Guide: Dementia Care
A careful, honest account of what social and companion robots can and cannot do in dementia care, what the research evidence actually shows, and what care homes and families should ask before any deployment.
Dementia care is the area of social robotics that has the strongest and oldest evidence base. Research into robotic and social devices for people living with dementia predates the current interest in humanoid robots by more than a decade. That head start matters, because it means there is more to say with confidence about what works, in what circumstances, and for whom. It also means the limitations are well documented.
This guide covers the actual state of play in the UK: which devices are in use, what the peer-reviewed research shows, what the honest limitations are, and what care homes and families need to consider before these technologies are introduced into a dementia care setting.
Important note: this guide provides general information for care professionals and families. It is not clinical guidance. Decisions about care for people living with dementia should always involve qualified clinical and care professionals, and should give central weight to the person's own expressed preferences and best interests.
It is important to be clear about terminology, because the word "robot" is used to describe a wide range of things in this context. The devices actually deployed in UK dementia care settings today are not ambulatory humanoid robots. They are more accurately described as social robots or companion robots: stationary or slow-moving devices designed to provide interaction, emotional engagement, and sometimes basic communication support.
The best-known device in this category is PARO, a soft robotic seal developed in Japan. PARO responds to touch, sound, and light with movements and vocalisations. It was designed specifically for use with older adults, including those with dementia, to provide a calming, non-verbal form of interaction. PARO has been evaluated in clinical settings across several countries, including the UK, over more than a decade. It is CE marked as a medical device in a number of jurisdictions.
Other therapeutic companion devices exist, including robotic pets and companion robots designed to simulate the presence and responsiveness of a living animal without the practical challenges of pet care in a clinical setting.
Devices such as the GenieConnect unit from Service Robotics are screen-based companion robots that support video calls, conversation, reminders, and entertainment. These are deployed in UK care homes on a Robot-as-a-Service basis. For residents with dementia, these devices may be used to maintain family contact or provide reminders, though their utility depends on the individual's stage of dementia and their comfort with screen-based interaction.
None of the above devices perform physical care. They do not assist with mobility, personal care, medication administration, or clinical monitoring. They are not a substitute for human care staff. Full ambulatory humanoid robots of the kind depicted in recent media coverage are not deployed in UK care homes at any significant scale at the time of writing. The gap between media coverage of humanoid robots and the devices that are actually present in dementia care settings is substantial.
The evidence base for social robots in dementia care is more developed than for almost any other area of care robotics. The following is a summary of what the research literature broadly reports. Individual studies vary in methodology and sample size; the following reflects the weight of evidence rather than any single study.
A number of studies, including systematic reviews, have found that therapeutic companion robots such as PARO are associated with reductions in agitation and anxiety in some people living with dementia. The effect is not universal: some individuals engage positively, others do not respond or are indifferent, and a small number find the devices distressing. The evidence is strong enough to justify considering these devices as one tool in a person-centred care plan, but not strong enough to recommend them as a general intervention for all people with dementia.
Studies have reported improvements in social engagement, positive affect, and communication in some care home residents following interaction with companion robots. Again, results vary considerably between individuals and settings. The best outcomes are associated with structured, staff-supported interaction rather than simply placing a device with a resident and leaving them with it.
The evidence does not support claims that social robots slow cognitive decline, improve clinical outcomes for dementia, or provide comparable benefits to human interaction. These are important limits. The value of companion robots in dementia care is in providing a source of engagement and calm for individuals who may be difficult to reach through conventional interaction, during periods when staff time is constrained. It is a supplement to human care, not a treatment and not a replacement for it.
Many studies in this field have small sample sizes and relatively short follow-up periods. Larger and longer-term UK-specific studies would strengthen the evidence base considerably. This is an active area of research, not a settled one.
The National Institute for Health and Care Excellence (NICE) has not, at the time of writing, issued specific guidance recommending or approving social robots as a clinical intervention for dementia. Care providers should not present these devices to residents or families as clinically recommended treatments. Their appropriate framing is as supportive, non-pharmacological engagement tools, deployed within a broader person-centred care plan.
Several UK care organisations have piloted or deployed social robots in dementia care settings. Reported examples include Service Robotics' GenieConnect units deployed across a number of UK care homes on a RaaS model, and Cera's Genie AI droids placed with residents at a care home in Lewisham. Both of these deployments are reported in UK care trade press and provider announcements; independent verification of outcomes is not available to this site at the time of writing.
PARO units have been used in NHS-adjacent and care home settings in the UK, though widespread systematic deployment has been limited partly by cost and partly by variable awareness among care managers of the evidence base.
The honest assessment is that UK deployment of social robots in dementia care is at a relatively early and uneven stage. A small number of progressive providers are actively piloting and deploying these technologies. The majority of UK dementia care settings have not yet introduced them. This is a market that is developing, not one that is mature.
Deploying any technology with people living with dementia raises specific ethical questions that go beyond the standard considerations for technology deployment in care settings. These are not reasons to avoid social robots; they are reasons to deploy them thoughtfully.
A person living with dementia may lack the capacity to consent to the use of a companion robot in their care. In such cases, the Mental Capacity Act 2005 requires that decisions are made in the person's best interests, taking into account their known preferences, values, and wishes. The involvement of family members and close carers in this assessment is important. The use of a companion robot with a person who has clearly indicated, at any stage of their illness, that they do not want such contact must be carefully considered against the best-interests framework.
Some practitioners raise a concern about "robot deception": whether it is ethically acceptable for a person with dementia to form an attachment to a device that is not a living being. This is a genuine ethical debate within the care and philosophy of medicine literature. There is no settled consensus. The most defensible position is that if a device provides genuine comfort and reduces distress for a person who lacks insight into its nature, and if their family and care team agree it is in their best interests, the engagement is ethically justifiable. Care providers should document this reasoning rather than deploying the device without ethical consideration.
Companion devices that can record audio or video, or that transmit data to a cloud platform, raise data protection considerations under UK GDPR. Data about a person's health condition and care status is special-category data and requires explicit consent or another lawful basis. Care providers must understand what data a device collects, where it is stored, who can access it, and what the supplier's data retention and deletion policies are before deploying it with residents.
There is a legitimate concern that companion robots may be used to justify reduced staffing rather than as a genuine supplement to human care. A care home that deploys companion robots and simultaneously reduces the number of care staff hours has not improved care. The CQC's assessment of staffing adequacy is independent of whether technology is in use; a well-led care home will use both appropriately.
If you are a care home manager or care director considering social robots for your dementia care setting, the following questions are worth addressing before any deployment.
A pilot with a small number of willing and consenting residents, with clear outcome measures and a defined review period, is a sound approach. It allows the care home to gather real evidence of benefit or otherwise in its own setting, rather than relying solely on external studies or supplier claims.
If your relative's care home has introduced, or is considering introducing, a companion robot in the dementia care unit, you are entitled to ask the care home for a clear explanation of what the device does, how consent is handled, and what data it collects. A good care home will welcome this conversation.
If your relative has expressed preferences about this kind of technology at any point when they had capacity to do so, those preferences are relevant to the best-interests assessment and should be communicated to the care team. If you hold lasting power of attorney for health and welfare, you have a formal role in best-interests decisions of this kind.
The honest assessment is that for some people living with dementia, companion robots can provide genuine comfort and a source of engagement. For others they are unhelpful or irrelevant. The individual matters. No blanket statement about whether companion robots are good or bad for people with dementia holds for every person in every situation.
If you have concerns about how the technology is being used with your relative, raise them with the care home manager first. If you remain concerned, the Care Quality Commission (cqc.org.uk) can be contacted with concerns about care standards in England. Equivalent regulators operate in Wales (Care Inspectorate Wales), Scotland (Care Inspectorate) and Northern Ireland (RQIA).
If you are a care provider researching companion robots for dementia care, or a family member looking for information about a specific deployment, write to us at hello@humanoidrobotcare.co.uk. We provide information only and are not a regulated clinical or care service.
Our For Providers page covers the broader operational and regulatory context for care homes evaluating robot technology. Our guide What robots actually do in UK care homes today provides essential background on the types of device currently deployed.