Key takeaways. Each line stands on its own. Every figure is repeated further down with its measure, sample size and date.
- 33% of U.S. adults aged 50 to 80 said they lacked companionship some of the time or often in 2024 (University of Michigan National Poll on Healthy Aging, 2024).
- 40% of U.S. adults 45 and older scored as lonely in 2025, up from 35% in 2010 and 2018 (AARP, 2025).
- Among U.S. adults 45 and older, loneliness fell with age, from 49% at 45 to 49 to 28% at 70 and older (AARP, 2025).
- Across nine long-term studies, loneliness dipped in midlife and rose again in older adulthood (Psychological Science, 2024).
- 29% of U.S. adults 50 to 80 felt isolated from others in 2024, down from 56% early in the pandemic (University of Michigan, 2024).
- Nearly 3 in 10 U.S. adults 65 and older lived alone in 2022, including 43% of women 75 and older (U.S. Census Bureau, 2024).
- 75% of U.S. adults 50 to 80 in fair or poor mental health said they lacked companionship (University of Michigan, 2024).
- 63% of U.S. adults 45 and older with household incomes under $25,000 were lonely, about twice the rate at $100,000 or more (AARP, 2025).
- Loneliness was linked to a 26% higher likelihood of early death, and social isolation to 29% (Holt-Lunstad et al., 2015).
- Loneliness was linked to a 31% higher risk of dementia across 608,561 people (Nature Mental Health, 2024).
- Poor social relationships were linked to a 29% higher risk of heart disease and a 32% higher risk of stroke (Heart, 2016).
- The mortality impact of social disconnection is similar to smoking up to 15 cigarettes a day (U.S. Surgeon General, 2023).
- About 1 in 6 people worldwide experience loneliness, linked to an estimated 871,000 deaths a year (World Health Organization, 2025).
- 17% of people 65 and older worldwide felt very or fairly lonely, against 27% of adults 19 to 29 (Meta-Gallup, 2023).
- Of 22 randomized trials testing AI chatbots, voice assistants or robots against loneliness, 12 were run with older adults: 11 tested robots, voice assistants or an on-screen agent, and one, published in October 2026, a chatbot companion (My Softly research map of 234 studies, November 2026).
A note on who compiled this: we make My Softly, an AI companion app whose paying subscribers are mostly in the age group this page covers. None of the figures here are ours. Every one comes from a named survey, study or government report, linked in the Sources below, and we checked each on the publisher's own page.
Surveys measure loneliness in different ways, so their numbers are not interchangeable. The University of Michigan poll counts people who lacked companionship "some of the time or often". AARP scores the 20-item UCLA Loneliness Scale. The CDC asks a single question in a telephone survey. UK and Canadian figures mostly count feeling lonely "often or always", which is a higher bar and gives lower numbers. Every table below names the measure, and it is worth naming it when you quote a figure.
How many older adults are lonely?
About a third of Americans aged 50 to 80 say they lack companionship at least some of the time: 33 percent in 2024, according to the University of Michigan's National Poll on Healthy Aging. A broader measure puts it higher, at 40 percent of adults 45 and older on AARP's 2025 survey.
| Statistic | Figure | Source | Sample and measure | Date |
|---|---|---|---|---|
| U.S. adults 50 to 80 who lacked companionship some of the time or often | 33% in 2024 (34% in 2018, as high as 42% in between) | University of Michigan National Poll on Healthy Aging, in JAMA | About 2,000 to 2,600 per wave (2,576 in 2024), new sample each year | 2018 to 2024 |
| U.S. adults 50 to 80 who felt isolated from others some of the time or often | 29% in 2024 (27% in 2018, 56% in 2020) | University of Michigan National Poll on Healthy Aging, in JAMA | About 2,000 to 2,600 per wave | 2018 to 2024 |
| U.S. adults 50 to 80 in contact with people outside their home once a week or less | 33% (28% in 2018, 46% in 2020) | University of Michigan National Poll on Healthy Aging | 2,563, NORC AmeriSpeak panel | Jan 2023 |
| U.S. adults 45 and older who scored as lonely | 40% in 2025 (35% in 2010 and 2018) | AARP / Ipsos | 3,276, UCLA Loneliness Scale | Aug 2025 |
| U.S. adults 45 and older estimated to be lonely | About 57.9 million (47.8 million in 2018) | AARP / Ipsos | Projection from the same survey | 2025 |
| U.S. adults 45 and older who say they feel lonely always or sometimes | 33% (33% in 2010, 31% in 2018) | AARP / Ipsos | 3,276, self-rated | Aug 2025 |
| Of those, have felt this way for more than 10 years | 26% (38% for six years or more) | AARP / Ipsos | 3,276, self-rated | Aug 2025 |
| U.S. adults 65 and older who lacked the social and emotional support they need | 19.1% (24.1% of all adults) | CDC, MMWR | BRFSS telephone survey, 26 states, 218,915 adults | 2022 |
| Community-dwelling U.S. adults 65 and older who were socially isolated | 24%, about 7.7 million (4% severely) | Cudjoe et al., Journals of Gerontology | NHATS Medicare-based sample; living arrangement, network size, religious attendance and social participation | 2011 data |
The pandemic spike has mostly unwound. Loneliness and isolation among adults 50 to 80 rose sharply in 2020 and were back close to 2018 levels by 2024, except among those aged 50 to 64, where rates had not returned to where they were. AARP's longer series shows a different trend on a different measure: a rise from 35 to 40 percent between 2018 and 2025 among adults 45 and older. The 24 percent isolation figure, which the National Academies' 2020 report used and many reports still repeat, describes 2011.
Does loneliness increase with age?
Not in a straight line. Across nine long-term studies, loneliness fell from young adulthood to midlife and rose again in older adulthood, while U.S. surveys that stop at 80 find adults in their fifties and early sixties lonelier than those in their late sixties and seventies.
| Statistic | Figure | Source | Sample and measure | Date |
|---|---|---|---|---|
| Loneliness across the lifespan | U-shaped: falls from young adulthood to midlife, rises again in older adulthood | Graham et al., Psychological Science | 128,118 people aged 13 to 103 in nine longitudinal studies, 20+ countries, pre-2020 data | 2024 |
| U.S. adults who scored as lonely, by age | 45 to 49: 49% 50 to 59: 45% 60 to 69: 41% 70+: 28% | AARP / Ipsos | 3,276, UCLA Loneliness Scale | Aug 2025 |
| U.S. adults who lacked companionship, by age | 50 to 64: 40% 65 to 80: 34% | University of Michigan National Poll on Healthy Aging | 2,563 | Jan 2023 |
| U.S. adults who felt isolated, by age | 50 to 64: 37% 65 to 80: 29% | University of Michigan National Poll on Healthy Aging | 2,563 | Jan 2023 |
| U.S. adults who reported loneliness, by age | 18 to 34: 43.3% 35 to 49: 31.9% 50 to 64: 27.8% 65+: 23.8% | CDC, MMWR | BRFSS, 26 states | 2022 |
| Canadians who feel lonely always or often, by age | 65 to 74: 9% 75+: 14% | Statistics Canada | Canadian Social Survey | Aug to Sept 2021 |
| People worldwide who feel very or fairly lonely, by age | 19 to 29: 27% 65+: 17% All ages: 24% | Meta-Gallup | About 1,000 people per country, 142 countries | June 2022 to Feb 2023 |
The findings fit together once you notice which ages each survey can see. The American surveys find loneliness falling from the late forties onward: AARP's rate drops from 49 percent at 45 to 49 to 28 percent at 70 and older, and adults 50 to 64 are lonelier than adults 65 to 80 in every year the Michigan poll has run. The long-term international studies, and Canada's split at 75, look further into old age, where loneliness rises again. So "older adults are the loneliest group" is not what the data says, and neither is "loneliness fades with age". The peak years differ by measure, and the oldest adults are the least surveyed.
How many older adults live alone?
Nearly 3 in 10 Americans aged 65 and older lived alone in 2022, according to the U.S. Census Bureau, and the share climbs steeply with age, reaching 43 percent of women 75 and older.
| Statistic | Figure | Source | Sample and measure | Date |
|---|---|---|---|---|
| U.S. adults 65 and older living alone | Nearly 3 in 10 | U.S. Census Bureau | Census living-arrangements data, private households | 2022 |
| U.S. women living alone, by age | 65 to 74: 27% 75+: 43% | U.S. Census Bureau | Census living-arrangements data | 2022 |
| U.S. adults 65 and older living alone, counting nursing homes and group quarters in the base | 26% in 2023 (29% in 1990) | Pew Research Center analysis of Census data | 1990 census, 2023 American Community Survey | 2023 |
| U.S. adults 65 and older living alone, by sex | Women: 31% (38% in 1990) Men: 19% (15% in 1990) | Pew Research Center | 2023 American Community Survey | 2023 |
| U.S. adults living alone, by age | 65 to 84: 24% 85+: 38% | Pew Research Center | 2023 American Community Survey | 2023 |
| U.S. adults 65 and older in poverty: living alone vs living with others | 1 in 5 vs 6% | Pew Research Center | 2023 American Community Survey | 2023 |
| U.S. adults 50 to 80 who lacked companionship: living alone vs living with others | 47% vs 33% | University of Michigan National Poll on Healthy Aging | 2,563 | Jan 2023 |
The Census and Pew figures differ because Pew includes the roughly 3 percent of older adults in nursing homes and other group quarters in its base. Pew's long view is worth knowing: the share of older Americans living alone has fallen since 1990, driven by women, while it has risen for men. And living alone is not the same as being lonely. In 2023 the Michigan poll found adults living alone far more likely to lack companionship, but by 2024 the rate among people living alone was lower than among those living with others.
How much time do older adults spend alone?
Americans aged 60 and older spent about seven hours a day alone on average, and more than 10 hours if they lived by themselves, according to a 2019 Pew Research Center analysis of federal time-use diaries from 2014 to 2017.
| Statistic | Figure | Source | Sample and measure | Date |
|---|---|---|---|---|
| Daily time alone, U.S. adults 60 and older | About 7 hours; more than 10 hours for those living alone | Pew Research Center analysis of the American Time Use Survey | Bureau of Labor Statistics time-use diaries | 2014 to 2017 data |
| Share who spent all of their measured time alone | 60+: 14% Under 60: 8% | Pew Research Center analysis of the American Time Use Survey | Bureau of Labor Statistics time-use diaries | 2014 to 2017 data |
| Daily hours alone, U.S. adults 45 and older | 5.6 on average; 7.3 among those who scored as lonely | AARP / Ipsos | 3,276, self-reported | Aug 2025 |
Who is most likely to be lonely later in life?
Loneliness after 50 tracks health, money, work and marital status more closely than age: it is highest among people in fair or poor health, on low incomes, out of work, or who have never married.
| Statistic | Figure | Source | Sample and measure | Date |
|---|---|---|---|---|
| U.S. adults 50 to 80 in fair or poor mental health | Lacked companionship: 75% (74% in 2018) Felt isolated: 77% | University of Michigan National Poll on Healthy Aging | 2024 wave | 2024 |
| U.S. adults 50 to 80 in fair or poor physical health | Lacked companionship: 53% (50% in 2018) Felt isolated: 52% (43% in 2018) | University of Michigan National Poll on Healthy Aging | 2024 wave | 2024 |
| U.S. adults 50 to 80 not working or receiving disability who lacked companionship | 52% (38% in 2018) | University of Michigan National Poll on Healthy Aging | 2024 wave | 2024 |
| U.S. adults 45 and older who scored as lonely, by household income | Under $25,000: 63% $25,000 to $74,999: 41% $75,000 to $99,999: 39% $100,000+: 32% | AARP / Ipsos | 3,276, UCLA Loneliness Scale | Aug 2025 |
| U.S. adults 45 and older who scored as lonely, by marital status | Never married: 62% Divorced or separated: 47% Widowed: 41% Married: 34% | AARP / Ipsos | 3,276, UCLA Loneliness Scale | Aug 2025 |
| U.S. adults 45 and older who scored as lonely, by work status | Not working: 57% Working: 40% Retired: 34% | AARP / Ipsos | 3,276, UCLA Loneliness Scale | Aug 2025 |
| Retirees who stay in touch with former colleagues | 41% of all retirees; 28% of retirees who scored as lonely | AARP / Ipsos | 3,276 | Aug 2025 |
| U.S. adults 45 and older who scored as lonely, by sex | Men: 42% Women: 37% (34% and 36% in 2018) | AARP / Ipsos | 3,276, UCLA Loneliness Scale | Aug 2025 |
| U.S. adults 50 to 80 who lacked companionship, by sex | Women: 40% Men: 34% | University of Michigan National Poll on Healthy Aging | 2,563 | Jan 2023 |
| Canadians 65 and older who are lonely, by marital status | Separated or divorced: 32% Widowed: 31% Never married: 29% Married or common-law: 13% | Statistics Canada | Canadian Health Survey on Seniors | 2019/2020 data |
| U.S. adults 45 and older with fewer friends than five years earlier | 29% in 2025 (25% in 2018, 23% in 2010) | AARP / Ipsos | 3,276 | Aug 2025 |
| U.S. adults 60 and older, community ties, 2010 vs 2025 | Volunteered in the past year: 47% to 33% Religious services monthly: 50% to 37% Belong to a community group: 32% to 25% | AARP / Ipsos | 2010, 2018 and 2025 surveys | Aug 2025 |
The two big U.S. surveys disagree about sex. AARP now finds men more likely to be lonely than women, a change since 2018, while the Michigan poll found women more likely to lack companionship. They use different measures, age ranges and years, so the honest summary is that the gap is small and its direction depends on how you ask. Retirement, on these numbers, is not the risk factor it is often assumed to be: retirees score lower than people still working and far lower than people out of work for other reasons. Our guide Retired and lonely: why it happens and what actually helps covers why the first months after work can still feel quiet.
Put together, these surveys line up with the risk factors in the National Academies' 2020 review of social isolation and loneliness in older adults: living alone, the loss of a spouse or other family, chronic health conditions, hearing and vision loss, limited mobility and cognitive decline. The surveys add low income and being out of work for reasons other than retirement. Several of these factors go together, which is why isolation and loneliness tend to cluster in the same people rather than spread evenly.
That review was written for the U.S. health care system. It recommended that clinicians periodically assess older patients for isolation and loneliness with validated tools, the way they check for other conditions, and connect those at risk with services in their community. As the research summarized in this article has shown, the health effects are large enough to justify the question.
What are the health effects of loneliness and isolation?
Large observational studies link loneliness and social isolation to earlier death, heart disease, stroke and dementia, with loneliness on its own linked to a 26 percent higher likelihood of death and a 31 percent higher risk of dementia.
| Statistic | Figure | Source | Sample and design | Date |
|---|---|---|---|---|
| Higher likelihood of death linked to loneliness / social isolation / living alone | 26% / 29% / 32% | Holt-Lunstad et al., Perspectives on Psychological Science | Meta-analysis of studies from 1980 to 2014 | 2015 |
| Greater likelihood of survival for people with stronger social relationships | 50%, over an average of 7.5 years | Holt-Lunstad et al., PLOS Medicine | Meta-analysis, 148 studies, 308,849 people | 2010 |
| Mortality impact of being socially disconnected | Similar to smoking up to 15 cigarettes a day | U.S. Surgeon General's advisory | Evidence review | 2023 |
| Higher risk of coronary heart disease / stroke with poor social relationships | 29% / 32% | Valtorta et al., Heart | 23 papers from 16 longitudinal datasets, 3 to 21 years of follow-up | 2016 |
| U.S. adults over 60 who died within six years: reported loneliness vs did not | 22.8% vs 14.2% | Perissinotto et al., Archives of Internal Medicine | Health and Retirement Study, 1,604 adults, mean age 71 | 2012 (2002 to 2008 data) |
| Higher risk of dementia linked to loneliness | 31% all-cause (39% Alzheimer's, 74% vascular dementia) | Luchetti et al., Nature Mental Health | Meta-analysis, 21 samples, 608,561 people | 2024 |
| Higher risk of dementia linked to social isolation | About 50% | National Academies of Sciences, Engineering, and Medicine | Consensus report | 2020 |
| Heart failure patients reporting high perceived social isolation: risk of death / hospitalisation / emergency visits | 3.74 times / 68% higher / 57% higher | Manemann et al., Journal of the American Heart Association | 1,681 patients in Minnesota, mean age 73 | 2018 |
| Extra Medicare spending linked to social isolation among older adults | $6.7 billion a year (about 4 million isolated enrollees, 14%) | AARP Public Policy Institute | Health and Retirement Study linked to Medicare claims | 2017 |
Three cautions before quoting these. They are associations from observational studies, not proof that loneliness causes illness. The "15 cigarettes a day" comparison describes social connection as a whole, not loneliness alone; for loneliness specifically, the same research team's estimate is the 26 percent higher likelihood of death in the first row. And the 50 percent dementia figure is about social isolation: the largest loneliness-only estimate is 31 percent. The 2015 meta-analysis also found that social deficits predicted death more strongly in samples with an average age under 65, not over it.
How common is loneliness among older adults worldwide?
About 1 in 6 people worldwide experience loneliness, according to the World Health Organization, which also estimates that up to 1 in 3 older adults are socially isolated. Countries that count only frequent loneliness report much lower figures for older people.
| Statistic | Figure | Source | Sample and measure | Date |
|---|---|---|---|---|
| People worldwide who feel lonely | About 1 in 6 | World Health Organization Commission on Social Connection | Pooled survey estimates, 2014 to 2023 | June 2025 |
| Deaths a year linked to loneliness, worldwide | About 871,000 (about 100 an hour) | World Health Organization Commission on Social Connection | Estimates for 2014 to 2019 | June 2025 |
| Older adults worldwide who are socially isolated | Up to 1 in 3 | World Health Organization Commission on Social Connection | Studies from 1990 to 2022 | June 2025 |
| Adults in England who feel lonely often or always, by age | 65 to 74: 4% 75+: 5% All adults: 7% | UK Government, Community and Engagement Survey | More than 175,000 responses | Oct 2025 to Mar 2026 |
| UK adults 65 and older who are often lonely | 7%, about 940,000 people | Age UK | Analysis of Understanding Society and the English Longitudinal Study of Ageing | Dec 2024 |
| Adults 65 and older in England who go a week without speaking to a friend or family member | 3%, about 270,000 people | Age UK | Same analysis | Dec 2024 |
| Canadians 65 and older who experienced loneliness | 19.2%, about 1.1 million (women more likely than men) | Statistics Canada, Health Reports | 38,941 adults aged 65 and older | 2019/2020 data |
Do older adults use technology or AI for company?
Most older Americans have the tools and very few use AI for company yet: 78 percent of adults 65 and older own a smartphone and 23 percent have used AI chatbots, but 1 percent say they have used chatbots for companionship.
| Statistic | Figure | Source | Sample and measure | Date |
|---|---|---|---|---|
| U.S. adults who own a smartphone, by age | 50 to 64: 90% 65+: 78% | Pew Research Center | 5,022 adults | 2025 |
| U.S. adults 65 and older with home broadband | 70% | Pew Research Center | 5,022 adults | 2025 |
| U.S. adults 50 and older who have used generative AI | 2023: 9% 2024: 18% 2025: 30% | AARP | 3,083 adults 50+ | Sept to Oct 2025 |
| U.S. adults who have ever used AI chatbots, by age | Under 50: 63% 50 to 64: about 4 in 10 65+: 23% | Pew Research Center | 5,119 adults | Feb 2026 |
| U.S. adults who have ever used chatbots for emotional support or advice / for companionship, by age | 50 to 64: 4% / 2% 65+: 2% / 1% | Pew Research Center | 5,119 adults | Feb 2026 |
| Adults 50 and older who rate AI for emotional support extremely or very interesting | 38% (45% in their fifties, 36% in their sixties, 28% at 70+) | AARP (opt-in panel) | 1,148 adults 50+ | Jan 2025 |
| U.S. adults 45 and older who scored as lonely and are interested in AI for companionship, conversation and reminders to connect | 23% | AARP / Ipsos | 3,276, UCLA Loneliness Scale | Aug 2025 |
Interest runs well ahead of use, and it is highest among people in their fifties, who also report more loneliness than older age groups in the surveys above. For how many people already use AI companions, by age and in total, see our AI companion statistics for 2026. For which apps suit people in their fifties and sixties, and what to look for, see AI companions for people over 50.
What helps with loneliness in older adults?
Regular, structured contact helps measurably: in a randomized trial, four weeks of empathy-focused phone calls from trained volunteers lowered loneliness on the UCLA scale among Meals on Wheels clients, and a review of AI tools tested with adults 55 and older found significant reductions in six of nine studies.
| Statistic | Figure | Source | Sample and design | Date |
|---|---|---|---|---|
| Change in loneliness after four weeks of empathy-focused phone calls from trained lay callers | 1.1-point improvement on the UCLA Loneliness Scale (Cohen's d 0.48) | Kahlon et al., JAMA Psychiatry | Randomized trial, 240 Meals on Wheels clients (63% aged 65+, 56% living alone) | 2021 |
| Studies of AI tools with adults 55 and older that found significant reductions in loneliness | 6 of 9 | Yang et al., Healthcare | Systematic review, six randomized trials and three pre-post studies, mostly social robots | 2025 |
| Change in loneliness among older people using conversational agents, before versus after | Mild to moderate drop (standardized mean change 0.35) | Satake et al., Psychological Medicine | Meta-analysis of 17 studies (9 robot, 8 non-robot), pooling before-and-after change rather than comparing with control groups | 2026 |
| Effect of AI conversational and socially assistive agents on loneliness in randomized trials with older adults | No significant effect on loneliness; small drop in depressive symptoms | Gou et al., BMC Geriatrics | Meta-analysis, 8 randomized trials, 611 older adults; the loneliness result pools 3 of them (190 people) | 2026 |
| Effect of social robots on loneliness in older adults | Significant reduction, larger in care homes than among people living independently | Mehrabi and Ghezelbash, The Gerontologist | Meta-analysis, 19 studies, 1,083 older adults | 2025 |
The phone-call trial is striking for how light the intervention was: 16 callers aged 17 to 23, calls every day for five days and then at least twice a week. What it supplied was regular contact from someone who asked and listened. Our guide to AI companions and loneliness goes through the research on where apps fit.
Has AI been tested against loneliness in older adults?
Robots and voice assistants many times; a chatbot companion once. We keep a running map of the research on AI companions, chatbots, voice assistants and social robots that measured loneliness or wellbeing: 234 studies, checked to November 6, 2026. 22 were randomized trials that measured loneliness. 12 of those were run with older adults: 7 tested social robots, 4 voice assistants or an on-screen agent, and one, published on October 23, 2026, tested a chatbot companion.
That trial, published in the Journal of Visualized Experiments, randomized 160 adults aged 60 and over in China who were at risk of social isolation to eight weeks with a GPT-4 virtual companion they were encouraged to talk to every day, by text or voice, or to their usual social activities. Loneliness on the UCLA scale fell from 52.3 to 38.6 in the companion group and from 51.8 to 49.5 in the comparison group, and the gap was still there four weeks later. It is one trial with clear limits: the assessors knew who was in which group, only the 148 people who finished were analysed, the paper reports no trial registration, and the drop is far larger than recent meta-analyses have found for any kind of agent. Until someone repeats it, it is a first result rather than a settled one.
| Type of AI | Studies found | With an older-adult sample | Randomized trials measuring loneliness | Of those, with older adults |
|---|---|---|---|---|
| Chatbots and companion apps | 118 | 9 | 9 | 1 |
| of which companion apps | 49 | 2 | 4 | 1 |
| Social robots and robot pets | 61 | 54 | 8 | 7 |
| Voice assistants and purpose-built eldercare agents | 26 | 22 | 5 | 4 |
| Reviews pooling several types | 29 | 20 | 0 | 0 |
| All | 234 | 105 | 22 | 12 |
A count shows how much research exists, not whether the thing works. On that question, a 2026 meta-analysis in Psychological Medicine found modest drops in loneliness among older people using conversational agents, measured before and after use rather than against a control group, while a 2026 meta-analysis in BMC Geriatrics found no significant effect on loneliness across the three randomized trials with older adults that measured it. Across the 9 chatbot trials that measured loneliness, 5 found a drop, 2 mixed results and 2 no change; apart from the one above, they enrolled students and adults averaging around 40 or younger and ran eight weeks at most. Our guide to AI companions and loneliness sets out the key trials and how sure each finding is.
The research map: download it, reuse it
Every study is in a spreadsheet: download the research map (CSV, 234 studies). Each row gives the study's design, what kind of AI it tested, who took part and how many, the country, the outcome measures, the direction of the finding and a link to the paper. It is free to reuse under CC BY 4.0 with a link to this page. Where a study found no effect, mixed results or harm for a named commercial product, the row gives the study, its design and the direction of the finding but not the product name or a one-line summary; the linked paper has both.
We search for new studies every quarter; the next update is due in February 2027. What changed so far:
- November 6, 2026 (234 studies). 55 studies added: 9 published since October 10 and 46 earlier ones the first search missed. Among them is the first randomized trial of a chatbot companion against loneliness in people over 60 (Yao, JoVE, October 23, 2026). Two earlier records were corrected after reading the full papers, and three chatbot studies for older adults were regrouped to match our definitions. A random tenth of the new records was rechecked against the sources.
- October 10, 2026 (179 studies). First edition: 179 studies from 45 OpenAlex searches (2016 onward), PubMed searches for trials, and the reference lists of four recent reviews. A random tenth rechecked against the sources.
How we counted. OpenAlex (45 queries, 2016 onward, rerun at each update) and PubMed searches, plus the reference lists of ten recent reviews; each study read from its abstract or full text; a random tenth of each update's new records rechecked against the source by a second reader. Included: any empirical study or review of an AI companion, chatbot, voice assistant, eldercare agent or social robot with a loneliness, connection, wellbeing or mood outcome; one row per trial. Older-adult sample means the study enrolled or separately analysed people 60 or 65 and over, or described its sample as older adults. Last checked November 6, 2026.
Figures we left out, and why
- "Loneliness raises the risk of death by 50 percent." A misreading of the 2010 meta-analysis, which measured the survival advantage of strong social relationships overall, not loneliness.
- "43 percent of adults 60 and older are lonely." It traces to a real study, but its data are from 2002. The 2024 Michigan figure replaces it.
- "Cognitive abilities decline 20 percent faster in lonely people." We could not find the 20 percent figure in the abstract of the study the Surgeon General's advisory cites for it.
- "1.4 million chronically lonely older people in England." Not on any current Age UK page; Age UK's December 2024 report uses 940,000 for the UK.
- Australian figures. We could not open the Australian Institute of Health and Welfare or Australian Bureau of Statistics pages to confirm the older-adult numbers that circulate, so none are included.
About My Softly
My Softly is our app: an AI companion app for iPhone and Android that remembers what you tell it and asks about your day, for the days nobody else does. The people paying for it are mostly in the age group this page is about: paying subscribers are in their mid-fifties on average, and most are women (September 2026), and the feature they ask for by name is being messaged first.
It is an AI companion built around lasting memory and check-ins that come first. With check-ins on, it messages first about something you actually mentioned, outside quiet hours you set. You choose whether it is a friend or a partner. $15.99 a month or $159.99 a year, for iPhone and Android in the US, Canada, the UK, Australia, New Zealand and Singapore, and free to download on the App Store and Google Play with starter messages. If you are comparing apps, start with the best AI companion apps in 2026 or the best AI companion apps for seniors.
How to cite this page
My Softly (2026). "Loneliness statistics for older adults (2026)." mysoftly.ai/guides/loneliness-statistics-older-adults.
The research map: My Softly (2026). "AI companions, chatbots, voice assistants, social robots and loneliness: research map" (dataset, checked to November 6, 2026). mysoftly.ai/data/ai-companion-loneliness-research-map.csv.
Please cite the original survey or study for each figure; every one is linked in the Sources below. If this page saved you the tracing, a link to it is appreciated. We recheck the sources whenever we update the page, and the date at the top reflects the last revision.
Sources
- Malani, P. N., et al. (2025). "Loneliness and Social Isolation Among US Older Adults." JAMA 333(3):254-257, from the University of Michigan National Poll on Healthy Aging. doi.org/10.1001/jama.2024.23213 (release, December 9, 2024: ihpi.umich.edu)
- Malani, P., Singer, D., Kirch, M., et al. (March 2023). "Trends in Loneliness Among Older Adults from 2018-2023." University of Michigan National Poll on Healthy Aging; NORC AmeriSpeak, n=2,563, January 2023. ihpi.umich.edu
- AARP Research (December 2025). "Disconnected: The escalating challenge of loneliness among adults 45-plus." Ipsos KnowledgePanel, n=3,276, August 4 to 19, 2025, 20-item UCLA Loneliness Scale. aarp.org
- Bruss, K. V., Seth, P., and Zhao, G. (June 20, 2024). "Loneliness, Lack of Social and Emotional Support, and Mental Health Issues, United States, 2022." Morbidity and Mortality Weekly Report 73(24), CDC. BRFSS, 26 states, n=218,915. pmc.ncbi.nlm.nih.gov
- Cudjoe, T. K. M., et al. (2020). "The Epidemiology of Social Isolation: National Health and Aging Trends Study." The Journals of Gerontology: Series B. Data from 2011. pubmed.ncbi.nlm.nih.gov
- Graham, E. K., et al. (April 30, 2024). "Do We Become More Lonely With Age? A Coordinated Data Analysis of Nine Longitudinal Studies." Psychological Science. 128,118 participants. doi.org/10.1177/09567976241242037
- Statistics Canada (November 24, 2021). "Canadian Social Survey: Loneliness in Canada." The Daily. statcan.gc.ca
- Statistics Canada (November 6, 2023). "A look at loneliness among seniors." StatsCAN Plus, from the Canadian Health Survey on Seniors 2019/2020. statcan.gc.ca
- Islam, M. K., and Gilmour, H. (July 19, 2023). "Immigrant status and loneliness among older Canadians." Health Reports, Statistics Canada. n=38,941 aged 65 and older. statcan.gc.ca
- Maese, E. (October 24, 2023). "Almost a Quarter of the World Feels Lonely." Gallup, from the Meta-Gallup Global State of Social Connections survey in 142 countries. news.gallup.com
- Hemez, P., and Washington, C. (May 30, 2024). "Living Arrangements Varied Across Age Groups." U.S. Census Bureau. census.gov
- Fry, R. (December 4, 2025). "A smaller share of older U.S. adults live alone today than in 1990." Pew Research Center. pewresearch.org
- Livingston, G. (July 3, 2019). "On average, older adults spend over half their waking hours alone." Pew Research Center, from American Time Use Survey data for 2014 to 2017. pewresearch.org
- Holt-Lunstad, J., Smith, T. B., Baker, M., Harris, T., and Stephenson, D. (2015). "Loneliness and social isolation as risk factors for mortality: a meta-analytic review." Perspectives on Psychological Science 10(2). pubmed.ncbi.nlm.nih.gov
- Holt-Lunstad, J., Smith, T. B., and Layton, J. B. (July 2010). "Social Relationships and Mortality Risk: A Meta-analytic Review." PLOS Medicine. 148 studies, 308,849 participants. journals.plos.org
- U.S. Surgeon General (May 2023). "Our Epidemic of Loneliness and Isolation." U.S. Department of Health and Human Services. ncbi.nlm.nih.gov
- Valtorta, N. K., et al. (2016). "Loneliness and social isolation as risk factors for coronary heart disease and stroke: systematic review and meta-analysis of longitudinal observational studies." Heart. pmc.ncbi.nlm.nih.gov
- Perissinotto, C. M., Stijacic Cenzer, I., and Covinsky, K. E. (2012). "Loneliness in older persons: a predictor of functional decline and death." Archives of Internal Medicine. Health and Retirement Study, n=1,604. pubmed.ncbi.nlm.nih.gov
- Luchetti, M., et al. (October 2024). "A meta-analysis of loneliness and risk of dementia using longitudinal data from >600,000 individuals." Nature Mental Health. pubmed.ncbi.nlm.nih.gov
- National Academies of Sciences, Engineering, and Medicine (2020). "Social Isolation and Loneliness in Older Adults: Opportunities for the Health Care System." ncbi.nlm.nih.gov
- Manemann, S. M., et al. (May 23, 2018). "Perceived Social Isolation and Outcomes in Patients With Heart Failure." Journal of the American Heart Association. n=1,681. pubmed.ncbi.nlm.nih.gov
- AARP Public Policy Institute (November 2017). "Medicare Spends More on Socially Isolated Older Adults." Insight on the Issues 125. aarp.org
- World Health Organization Commission on Social Connection (June 30, 2025). "From loneliness to social connection: charting a path to healthier societies." who.int (plain-language summary: PDF)
- UK Government (August 14, 2026). "Community and Engagement Survey 2025/26: loneliness report." More than 175,000 responses, October 2025 to March 2026. gov.uk
- Age UK (December 2, 2024). "Age UK's new report shows 'you are not alone in feeling lonely'." Analysis of Understanding Society and the English Longitudinal Study of Ageing. ageuk.org.uk
- McClain, C., and Bishop, W. (January 8, 2026). "What we know about internet use, smartphone ownership and digital divides in the U.S." Pew Research Center. pewresearch.org
- Gottfried, J., et al. (June 17, 2026). "How Americans' opinions and use of AI differ by age." Pew Research Center. Survey of 5,119 U.S. adults, February 2026. pewresearch.org
- AARP Research (December 2025). "2026 Tech Trends and the 50-Plus." n=3,083 aged 50 and older, September 9 to October 6, 2025. aarp.org
- AARP Research (September 2025). "Navigating the World of AI: Attitudes, Awareness and Openness Among Adults Ages 50-plus." Opt-in online panel, n=1,148 aged 50 and older, January 2025. aarp.org
- Kahlon, M. K., et al. (2021). "Effect of Layperson-Delivered, Empathy-Focused Program of Telephone Calls on Loneliness, Depression, and Anxiety Among Adults During the COVID-19 Pandemic: A Randomized Clinical Trial." JAMA Psychiatry 78(6). n=240. pmc.ncbi.nlm.nih.gov
- Yang, Y., Wang, C., Xiang, X., and An, R. (February 20, 2025). "AI Applications to Reduce Loneliness Among Older Adults: A Systematic Review of Effectiveness and Technologies." Healthcare 13(5):446. doi.org/10.3390/healthcare13050446
- Satake, Y., et al. (January 20, 2026). "Autonomous conversational agents for loneliness, social isolation, depression and anxiety in older people without cognitive impairment: Systematic review and meta-analysis." Psychological Medicine. doi.org/10.1017/s0033291725103073
- Yao, H. (October 23, 2026). "A Methodological Framework for a Large Language Model Virtual Companion for Older Adults at Risk of Social Isolation." Journal of Visualized Experiments. doi.org/10.3791/73774
- Gou et al. (2026). "Effectiveness of AI-based conversational and socially assistive agents in older adults: a systematic review and meta-analysis." BMC Geriatrics. doi.org/10.1186/s12877-026-07418-6
- Mehrabi and Ghezelbash (2025). "Wired for companionship: a meta-analysis on social robots filling the void of loneliness in later life." The Gerontologist. doi.org/10.1093/geront/gnaf219
Frequently asked questions
What percentage of older adults are lonely?
Between about a quarter and 40 percent, depending on the measure. The University of Michigan's National Poll on Healthy Aging found 33 percent of adults aged 50 to 80 lacked companionship some of the time or often in 2024. AARP's 2025 survey, using the 20-item UCLA Loneliness Scale, classed 40 percent of adults 45 and older as lonely. CDC data for 2022 found 23.8 percent of adults 65 and older reported loneliness.
Does loneliness get worse as you get older?
Not in a straight line. A 2024 analysis of nine long-term studies with 128,118 people found loneliness falls from young adulthood to midlife and rises again in older adulthood. In U.S. surveys, adults 50 to 64 report more loneliness than adults 65 to 80, and AARP found it fell to 28 percent at 70 and older. In Canada, adults 75 and older were more likely than those 65 to 74 to feel lonely always or often (14 versus 9 percent).
How many older adults live alone in the US?
Nearly 3 in 10 adults 65 and older lived alone in 2022, according to the U.S. Census Bureau, including 43 percent of women 75 and older. Pew Research Center's analysis of 2023 data, which also counts people in nursing homes, puts the share at 26 percent, down from 29 percent in 1990. Older women are more likely to live alone than older men (31 versus 19 percent).
Is loneliness as bad as smoking 15 cigarettes a day?
The U.S. Surgeon General's 2023 advisory says the mortality impact of being socially disconnected is similar to smoking up to 15 cigarettes a day. That comparison rests on a 2010 meta-analysis of 148 studies on social relationships as a whole, not on loneliness alone. For loneliness specifically, a 2015 meta-analysis found a 26 percent higher likelihood of early death.
How many older adults are socially isolated?
29 percent of adults aged 50 to 80 said they felt isolated from others at least some of the time in 2024, according to the University of Michigan. By a stricter measure based on living arrangements and social ties, 24 percent of community-dwelling adults 65 and older were socially isolated in 2011 national data. The World Health Organization estimates that up to 1 in 3 older adults worldwide are socially isolated.
What are the health effects of loneliness in older adults?
Poor social relationships are linked to a 29 percent higher risk of coronary heart disease and a 32 percent higher risk of stroke. A 2024 meta-analysis of 608,561 people linked loneliness to a 31 percent higher risk of dementia. Among U.S. adults over 60, those reporting loneliness were more likely to die within six years (22.8 versus 14.2 percent). All of these are associations from observational studies.
Who is most likely to be lonely later in life?
People in poor health, on low incomes, not working, or who have never married. In the University of Michigan poll, 75 percent of adults 50 to 80 in fair or poor mental health lacked companionship, as did 53 percent of those in fair or poor physical health. AARP found the highest rates among adults 45 and older who had never married (62 percent), earned under $25,000 (63 percent) or were not working (57 percent). Retirees were lower, at 34 percent.
Has an AI companion been tested against loneliness in older adults?
Once, in a randomized trial. Of 22 randomized trials of AI against loneliness in our research map (234 studies, checked to November 6, 2026), 12 were run with older adults. Eleven tested social robots, voice assistants or an on-screen agent. One, published in October 2026, tested a GPT-4 chatbot companion with 160 adults aged 60 and over in China and found a large drop in loneliness after eight weeks; it is a single trial with clear limits and has not yet been repeated.
Do older adults use AI or technology for company?
Most have the tools and few use them for company so far. 78 percent of Americans 65 and older own a smartphone and 23 percent have used AI chatbots, but 1 percent say they have used them for companionship and 2 percent for emotional support or advice (Pew Research Center, 2026). Interest is higher than use: 38 percent of adults 50 and older rated AI for emotional support extremely or very interesting (AARP, 2025). Our AI companion statistics page has the usage figures in full.