An interventional study of AI-based conversational intervention in Mild Cognitive Impairment (MCI) and Social Isolation, sponsored by Massachusetts General Hospital. Not yet recruiting at 1 site in United States. Open to participants aged 75 Years and older. Per ClinicalTrials.gov, last updated 2026-08-25.
Sponsored by Massachusetts General Hospital · Not applicable, Interventional, and Prevention
This randomized controlled trial examines the mechanisms of conversational AI-based interventions for early dementia prevention in socially isolated older adults.
Social isolation is a well-established and modifiable risk factor for cognitive decline and dementia. The completed I-CONECT trial (NCT02871921) showed that frequent, one-to-one, semi-structured conversational engagement delivered remotely by trained human interviewers improved cognitive outcomes among socially isolated older adults with mild cognitive impairment. Two limitations have constrained translating that approach into a scalable preventive intervention: its reliance on trained human interviewers limits scalability, and the behavioral mechanisms responsible for its cognitive benefits remain incompletely understood.
AI-CONECT addresses both limitations by delivering the conversational intervention through a conversational artificial intelligence (AI) voice chatbot rather than a human interviewer, and by testing a hypothesized behavioral mechanism: that the intervention increases participants' social self-efficacy (confidence in their ability to engage socially), which in turn increases real-world social engagement.
Exclusion Criteria:
Participants will receive four 15- to 30-minute conversations per week with the AI agent and a weekly 10- to15-minute survey call with a human interviewer.
Behavioral: AI-based conversational intervention
Participants will receive a weekly 10- to 15-minute survey call with a human interviewer.
AI-based one-on-one semi-structured conversation for cognitive stimulation
Social Self-Efficacy
Participants' self-efficacy on social interactions using an instrument adapted from the Exercise Self-Efficacy Scale (EXSE, by McAuley, 1993). The scale comprises three items of confidence: interacting with others regularly, engaging with others when feeling less connected, and engaging others in conversations. The three items are adapted to social interaction in person and using technology (e.g., video calls), respectively. Each item is rated from 0 (not at all confident) to 10 (completely confident). A mean score is computed across the six items (range 0-10), with higher scores indicating greater social self-efficacy.
Time frame: From enrollment to the end of intervention at 8 weeks
Time contacting friends or family
Change in weekly time contacting friends or family.
Time frame: From enrollment to the end of the intervention at 8 weeks, using weekly data
Time out of home
Self-reported weekly time out of home (a proximate measure of social engagement by the participants), using weekly data.
Time frame: From enrollment to the end of treatment at 8 weeks
Frequency of contacting friends
Self-reported frequency of contacting friends, using weekly data.
Time frame: From enrollment to the end of treatment at 8 weeks
Exploratory outcome: Social isolation score
Changes in Lubben's social isolation score (6-item version) from baseline to Week 8. The 6-item Lubben Social Network Scale (LSNS-6) is a brief measure of social isolation/social network size and support, particularly in older adults. It includes three questions about family ties and three about friendships, assessing the number of people the individual has contact with, feels close to, and can call on for help. Each item is scored from 0 to 5, giving a total score of 0-30. Higher scores indicate stronger social networks and less social isolation.
Time frame: From baseline to week 8
Exploratory outcome: Changes in the 3-item UCLA Loneliness Scale (UCLA-3)
The 3-item UCLA Loneliness Scale (UCLA-3) is a brief measure of perceived loneliness and social disconnection. It asks how often a person feels a lack of companionship, left out, and isolated from others. Each item is scored from 1 (hardly ever) to 3 (often). The total score ranges from 3 to 9, with higher scores indicating greater loneliness.
Time frame: From baseline to week 8
Plan to share: Yes — De-identified transcriptions and other collected data may be shared with external researchers under a Data Use Agreement (DUA). Shareable data may include outcome scores, adherence rates, chatbot-usage metrics, full conversational transcripts (after removal of any incidentally disclosed identifiable information), and audio recordings of sessions. Because raw audio recordings may be inherently identifiable, they will be shared only under a DUA. These data may also be made available to the NIH/NIA Program Official upon request. Data are stored on HIPAA-compliant servers (e.g., MGB Business Dropbox and MGB Azure) and encrypted drives. The participant linkage key is kept within Mass General Brigham. No individual-level data are deposited in public or controlled-access repositories under the current protocol; future repository sharing would require additional IRB review and approval.
Supporting information: Analytic code
No publications or documents are linked to this record.
This study is not yet recruiting, as verified in Aug 2026. You cannot join it, but the record below documents what was studied.
Get an email when the registry record changes — status, dates, results — or when someone posts here.
Sign in to followQuestions and observations about this study, from anyone following it. Not medical advice, and not a channel to the study team — their contact details are on the registry record.
Sign in to join the discussion. Reading takes no account; posting does. You choose a display name, and a pseudonym is the default.
Nothing here yet. If you are running this trial, taking part in it, or weighing whether to, this is the place to say so.
Massachusetts General Hospital