Guide 6
Loneliness among older people is a serious and well-documented public health concern in the UK. Social robots are one proposed response. This guide examines what the evidence actually supports, what it does not, and what families and care professionals should know before considering this technology.
Loneliness in older age is not a minor inconvenience. Research consistently links chronic loneliness to increased risk of poor physical health outcomes, cognitive decline, and earlier mortality. The UK has recognised this as a public health issue: a national strategy on loneliness exists, and care commissioners are under increasing pressure to address isolation as part of the wider social care agenda.
Social robots, companion devices, and AI-powered communication units are being trialled and deployed in the UK as one tool among many for addressing elder loneliness. The claims made for these technologies range from the careful and measured to the significantly overstated. This guide takes a measured position: the evidence base is growing and contains genuine positive findings, but it also contains important limitations that any responsible decision-maker needs to understand.
This guide provides general information only. It is not clinical or care advice. Decisions about care for specific individuals should involve qualified professionals.
Loneliness is distinct from simply being alone. A person can be physically isolated without feeling lonely, and a person surrounded by people can feel profoundly lonely. Research has consistently found that chronic loneliness is associated with measurably worse health outcomes. The mechanisms proposed include elevated stress hormones, disrupted sleep, reduced motivation for self-care, and lower engagement with health services.
In the UK context, loneliness among older people is particularly acute in several situations: following bereavement; after a move into residential care (when social networks built over a lifetime are disrupted); among people living alone with limited mobility; and among those living with dementia or other cognitive conditions that affect communication and memory.
The UK social care sector is not resourced to provide the level of human social contact that would resolve this problem at scale. Social care workers are doing demanding physical and emotional care work with limited time. Family visits vary enormously. There is a genuine, structural gap that technology is being asked, partly, to fill.
The category of "social robot" covers a wide range of devices. In the UK care context, the most relevant are companion communication devices: units that can hold conversations using AI, support video calls with family, play music, deliver reminders, and respond to voice commands. These are not humanoid robots in any science-fiction sense. They are, in practical terms, capable and interactive communications assistants in a physical form designed for the care context.
Some devices are designed specifically for social engagement. The PARO robotic seal, developed in Japan, is specifically designed for therapeutic use with older adults and people living with dementia. It responds to touch and voice, produces calming sounds, and has been the subject of more peer-reviewed research than almost any other device in this category.
AI-powered companion devices such as those used in UK care pilots by Cera and Service Robotics (GenieConnect) are different in character: they are primarily communication and conversation platforms. They can hold open-ended conversations, answer questions, play games, read news, and facilitate family video calls, including in formats designed to be accessible for people with limited dexterity or mild cognitive impairment.
What these devices do not do: they do not provide physical companionship, they cannot substitute for the depth of a long-term human relationship, and they do not resolve the structural causes of loneliness in care settings.
The research base for social robots and loneliness in older adults is growing but must be read carefully. Several points are worth making about the quality and limitations of the current evidence before summarising what it shows.
Most studies in this area are small-scale, conducted over relatively short periods, and use self-reported measures of loneliness or wellbeing that are inherently subjective. Many studies have not been replicated at scale. Publication bias (studies with positive findings are more likely to be published than those with null results) means the published literature may overstate the effectiveness of these interventions. These are not reasons to dismiss the evidence, but they are reasons not to overstate what it says.
Studies of PARO and similar therapeutic robots in care settings have found reductions in self-reported loneliness and anxiety in some populations, particularly among people living with dementia. These are genuine findings and represent a real potential benefit for a subset of residents. Some studies have reported reductions in the need for medication to manage agitation in dementia care, though these findings require replication and should not be taken as established clinical fact.
Research on AI companion devices and video-call-enabled robots has found that regular family contact via these devices can improve self-reported mood and reduce feelings of disconnection. Where the barrier to family contact has been the technology (difficulty using smartphones or tablets), a purpose-built device that is easier to use can make a measurable difference to contact frequency and quality.
There is limited high-quality evidence that any current social robot or companion device produces sustained, long-term reductions in loneliness as measured by validated scales. The most rigorous studies tend to show modest, time-limited effects that diminish once novelty wanes. There is no strong evidence that companion robots are as effective as structured human social activity for addressing loneliness in cognitively intact older adults.
Evidence on long-term outcomes, quality of life as measured by established instruments, or impact on physical health outcomes is limited. The research base is not yet at the level where these technologies can be recommended as evidence-based clinical interventions. They can be considered as one component of a broader approach to wellbeing.
If you are a care home manager or commissioner considering companion technology as part of a loneliness-reduction strategy, the following points are worth considering alongside the evidence.
Companion devices are most effective when deployed as a supplement to, not a replacement for, structured social activity, meaningful engagement with staff, and regular family contact. A care home with strong activities programming and engaged staff will see better outcomes from adding companion technology than one that deploys technology as a substitute for investment in human interaction.
Engagement with companion devices is highly variable between residents. Some older people engage readily and find them a source of genuine comfort and entertainment. Others find them uncomfortable, confusing, or simply not of interest. Any deployment must respect individual preferences. Residents should never be required to interact with a companion device, and their preferences and reactions should be recorded as part of their care plan.
For residents with mental capacity, consent to use a companion device should be sought and recorded. For residents who lack capacity, decisions should follow the Mental Capacity Act 2005 best-interests framework, involving family members and those who know the resident well. A companion device should not be deployed for a resident who lacks capacity simply because the home has purchased it; the individual's likely preferences, known history, and best interests must be the starting point.
The most effective deployments appear to be those where companion technology is integrated into the resident's care plan with specific goals identified: increasing family contact frequency, reducing daytime agitation, providing stimulation during periods when activities staff are unavailable. A device that is placed in a room and left without any plan for engagement is unlikely to deliver meaningful benefit.
If your relative is in a care home that uses companion robots, or if you are considering a device for a relative living at home, the key question is always whether the technology fits your relative as an individual. Ask: does your relative engage with it? Does it make them more connected to family or more entertained and comfortable? Or does it confuse or distress them?
If the answer to the last question is yes, that preference should be communicated to the care home and respected. Technology is not an end in itself.
If your relative lives alone and you are considering a companion device, be realistic about what it can offer. The strongest case for a companion device for a person living alone is that it enables easier, more frequent family contact in a format suited to your relative's abilities. It is not a substitute for regular human visits or for addressing other aspects of social isolation. If your relative is not already comfortable with technology, a companion device introduces a new learning curve that may itself be a source of frustration.
Social robots and companion devices show genuine but modest promise as one component of a broader approach to loneliness in older people. The evidence base supports cautious optimism, not strong clinical claims. For some residents, particularly those in dementia care and those with limited family contact due to technology barriers, these devices can provide a meaningful benefit. For others, they will simply not engage.
The most important thing a care home or family can do is approach this honestly: evaluate the technology against the individual, deploy it thoughtfully within a proper care plan, and measure whether it is actually helping. Technology that does not demonstrably benefit the individual should not be continued for its own sake.
If you have questions about companion technology for your care setting or for a relative, write to us at hello@humanoidrobotcare.co.uk. We provide information, not clinical advice.
Common questions
The evidence suggests that social robots and companion devices can reduce self-reported feelings of loneliness and isolation in some older people, particularly in care home settings and in those with dementia. However, effects are modest, variable between individuals, and often time-limited. The research base is growing but is not yet at the level where strong clinical claims are warranted. Companion devices work best as a supplement to human contact and structured social activity, not as a standalone solution.
The research shows genuine but modest positive findings. Studies of therapeutic devices such as PARO have found reductions in self-reported loneliness, anxiety, and agitation in some populations with dementia. Research on AI companion devices has found that enabling easier family contact can improve mood and reduce feelings of disconnection. However, most studies are small-scale, short-term, and use self-reported measures. There is limited evidence of sustained long-term effects, and no strong evidence that current devices are as effective as structured human social activity for cognitively intact older adults.