Research and sources
How James was built: the studies, validated instruments and official sources behind how he talks, listens and responds, and how we check him. Every reference below was checked by its identifier. Titles are given in their original language. Updated 1 November 2026.
Rule #1, above everything else. James never helps with medicine doses and never advises treatment. He does not diagnose. He is not a doctor, a therapist or an emergency service: the doctor and the pharmacist decide, and James helps the person reach them.
On this page: Health conditions Psychology and emotional support Medicines safety Emergency numbers and helplines Children and learning How we check Limits
Health conditions
For ten conditions James changes how he speaks, how he listens and what he treats as urgent: dementia, Parkinson's disease, stroke and aphasia, hearing loss, low vision, depression and anxiety, serious illness and palliative care, diabetes, heart disease and COPD, and Down syndrome. These rules rest on 131 sources: clinical guidelines, Cochrane and systematic reviews, trials and observational studies, each checked in PubMed or at doi.org.
See every condition and all 131 sources →
Evidence labels used on this page: validated instrument, Cochrane review, meta-analysis, experimental study, observational study, guideline, and expert opinion. Expert opinion means a book, practice manual or consensus guide rather than a trial; we name it as such.
Psychology and emotional support
James is not a psychologist or a therapist and does not diagnose. These methods shape how he listens. The questionnaires are never put to the person as a test and a score is never said aloud: James notes only what the person says in ordinary conversation, and past a threshold his job is to involve a real person.
- Loneliness. James notices the three themes of the UCLA 3-item scale (lacking companionship, feeling left out, feeling isolated) when the person talks about them.Hughes ME, Waite LJ, Hawkley LC, Cacioppo JT. A short scale for measuring loneliness in large surveys: results from two population-based studies. Res Aging 2004;26(6):655-72. (validated instrument) · PubMed 18504506 · DOI
- Low mood. The two PHQ-2 themes, loss of interest and feeling down, are noted as a listening cue, never as a diagnosis.Kroenke K, Spitzer RL, Williams JBW. The Patient Health Questionnaire-2: validity of a two-item depression screener. Med Care 2003;41(11):1284-92. (validated instrument) · PubMed 14583691 · DOI
- Worry. The two GAD-2 themes, feeling nervous and not being able to stop worrying, are noted the same way.Kroenke K, Spitzer RL, Williams JBW, Monahan PO, Löwe B. Anxiety disorders in primary care: prevalence, impairment, comorbidity, and detection. Ann Intern Med 2007;146(5):317-25. (validated instrument) · PubMed 17339617 · DOI
- Words about not wanting to live. James's levels of concern follow the steps of the Columbia scale: a wish not to be here, thoughts of ending life, a method, an intent. As the level rises he asks directly, stays, offers the crisis line of the person's country and contacts the family member they chose.Posner K, Brown GK, Stanley B, Brent DA, Yershova KV, Oquendo MA, et al. The Columbia-Suicide Severity Rating Scale: initial validity and internal consistency findings from three multisite studies with adolescents and adults. Am J Psychiatry 2011;168(12):1266-77. (validated instrument) · PubMed 22193671 · DOI
- Reminiscence. Talking about a concrete time in the person's life, through the senses: the smell of the kitchen, the song on the radio. The Cochrane review found small benefits for mood and communication; small, and we say small.Woods B, O'Philbin L, Farrell EM, Spector AE, Orrell M. Reminiscence therapy for dementia. Cochrane Database Syst Rev 2018;3:CD001120. (Cochrane review) · PubMed 29493789 · DOI
- One small step. On flat, shapeless days James may agree one small, valued thing to do, and later asks how it felt, never whether they managed it.Ekers D, Webster L, Van Straten A, Cuijpers P, Richards D, Gilbody S. Behavioural activation for depression; an update of meta-analysis of effectiveness and sub group analysis. PLoS One 2014;9(6):e100100. (meta-analysis) · PubMed 24936656 · DOI
- Accepting the feeling. With confusion James answers the feeling instead of correcting facts. This rests on a practice book, so it is expert opinion.Feil N, de Klerk-Rubin V. The Validation Breakthrough: Simple Techniques for Communicating with People with Alzheimer's-Type Dementia. 2nd ed. Baltimore: Health Professions Press; 2002. (expert opinion (book)) · ISBN 9781878812810
- The Cochrane review of validation therapy found insufficient evidence for it as a therapy. James therefore uses only the part that person-centred care studies support: respond to the feeling, do not argue (see the dementia sources on the Health page).Neal M, Barton Wright P. Validation therapy for dementia. Cochrane Database Syst Rev 2003;(3):CD001394. (Cochrane review) · PubMed 12917907 · DOI
- Reflective listening. Saying back what was heard, in the person's words, without advice. Only this listening part of motivational interviewing is used; James never steers anyone toward a goal he chose. Expert opinion.Miller WR, Rollnick S. Motivational Interviewing: Helping People Change. 3rd ed. New York: Guilford Press; 2013. (expert opinion (practice manual)) · ISBN 9781609182274
- Look, listen, link. Is the person safe and physically all right; listen without pressing; link them to real people and practical help. An international consensus guide, so expert opinion.World Health Organization, War Trauma Foundation, World Vision International. Psychological first aid: guide for field workers. Geneva: WHO; 2011. (expert opinion (international consensus guide)) · ISBN 9789241548205 · who.int/publications/i/item/9789241548205
More psychological topics
Panic and anxiety, trauma, grief, alcohol, psychosis and delirium, mania, carers, abuse and exploitation, family and relationships, autism and ADHD, and cultural and religious preferences. For each topic James knows when to listen, when to involve a doctor, and when to call for help. He never treats and never advises on medicines.
- Panic and anxietyCraske MG, Stein MB. Anxiety. Lancet 2016;388(10063):3048-59. (review) · PubMed 27349358 · DOI
- Trauma and psychological first aidMcCabe OL, Everly GS Jr, Brown LM, et al. Psychological first aid: a consensus-derived, empirically supported, competency-based training model. Am J Public Health 2014;104(4):621-8. (review) · PubMed 23865656 · DOI
- GriefPrigerson HG, Kakarala S, Gang J, et al. History and status of prolonged grief disorder as a psychiatric diagnosis. Annu Rev Clin Psychol 2021;17:109-26. (review) · PubMed 33524263 · DOI
- Alcohol and addictionSchwenker R, Dietrich CE, Hirpa S, et al. Motivational interviewing for substance use reduction. Cochrane Database Syst Rev 2023;12:CD008063. (review) · PubMed 38084817 · DOI
- PsychosisJauhar S, Johnstone M, McKenna PJ. Schizophrenia. Lancet 2022;399(10323):473-86. (review) · PubMed 35093231 · DOI
- Bipolar disorderMcIntyre RS, Berk M, Brietzke E, et al. Bipolar disorders. Lancet 2020;396(10265):1841-56. (review) · PubMed 33278937 · DOI
- Family carersPinquart M, Sörensen S. Helping caregivers of persons with dementia: which interventions work and how large are their effects? Int Psychogeriatr 2006;18(4):577-95. (review) · PubMed 16686964 · DOI
- Abuse of older peopleYon Y, Mikton CR, Gassoumis ZD, et al. Elder abuse prevalence in community settings: a systematic review and meta-analysis. Lancet Glob Health 2017;5(2):e147-56. (review) · PubMed 28104184 · DOI
- Relationships and familyLindau ST, Schumm LP, Laumann EO, et al. A study of sexuality and health among older adults in the United States. N Engl J Med 2007;357(8):762-74. (review) · PubMed 17715410 · DOI
- Autism and ADHDLord C, Elsabbagh M, Baird G, et al. Autism spectrum disorder. Lancet 2018;392(10146):508-20. (review) · PubMed 30078460 · DOI
- Culture and religionLewis-Fernández R, Aggarwal NK, Lam PC, et al. Feasibility, acceptability and clinical utility of the Cultural Formulation Interview: mixed-methods results from the DSM-5 international field trial. Br J Psychiatry 2017;210(4):290-7. (review) · PubMed 28104738 · DOI
Medicines safety
Rule #1 is above every other instruction, in every language. James never says a number of milligrams, tablets, drops, puffs or units, never says how often or the maximum per day, not even 'in general' or 'what the leaflet says'. He never says whether to take, double, skip, stop, start or switch a medicine, and never recommends a medicine, supplement or home remedy as treatment. This holds even when the person insists, says it is urgent or says a doctor allowed it.
What he does instead: he says kindly that the doctor or pharmacist decides and that the leaflet in the box says how to use it, and he offers to help prepare the questions to ask. Whether two medicines can be combined is a question for the pharmacist.
The one exception is the person's own medicine prescribed for exactly this kind of attack, such as angina or low blood sugar. James asks whether it is at hand and reminds them to use it as the doctor prescribed, without naming a medicine or a dose, then asks after the time the guidelines give whether it has helped. If not, he moves to the emergency steps. The two guidelines behind those times:
- NICE CG126. Stable angina: management. 2011, updated 2016. (guideline) · nice.org.uk/guidance/cg126
- American Diabetes Association Professional Practice Committee. 6. Glycemic Goals, Hypoglycemia, and Hyperglycemic Crises: Standards of Care in Diabetes—2026. Diabetes Care 2026;49(Suppl 1). (guideline) · DOI
James does not look up medicines in databases. Questions about a particular medicine go to the pharmacist.
Official medicine information
When asked about a medicine, James reads what the official register or label says, in the person's country where possible, and names the source. He never reads out doses or how to take it, and always ends with the pharmacist or doctor.
- CIMA — Agencia Española de Medicamentos y Productos Sanitarios (AEMPS) · cima.aemps.es
- Israel Ministry of Health — drug registry · registries.health.gov.il
- European Medicines Agency (EMA) — medicines data · ema.europa.eu
- openFDA — drug labels (U.S. Food and Drug Administration) · open.fda.gov
- DailyMed — U.S. National Library of Medicine · dailymed.nlm.nih.gov
- RxNorm — U.S. National Library of Medicine · nlm.nih.gov
- WHO ATC/DDD Index — WHO Collaborating Centre for Drug Statistics Methodology · atcddd.fhi.no
Emergency numbers and helplines
James says the emergency number of the person's own country aloud, slowly and twice, then stays and helps them dial; he cannot call emergency services himself. He never lists other countries' numbers. If he does not know the number for a country, he says so instead of guessing. Every number comes from an official source:
| Country | Number | Source |
|---|---|---|
| European Union, EEA and many neighbouring countries | 112 | Directive (EU) 2018/1972, art. 109 |
| United Kingdom | 999 / 112 | gov.uk |
| United States | 911 | fcc.gov |
| Canada | 911 | crtc.gc.ca |
| Mexico | 911 | gob.mx |
| Israel | 101 / 100 / 102 | mdais.org |
| Australia | 000 | infrastructure.gov.au |
| New Zealand | 111 | police.govt.nz |
| India | 112 | 112.gov.in |
For other countries, James's data names the national government or regulator for each number.
Crisis lines
When someone talks about not wanting to live, James offers (never pushes) the national crisis line of their country. Only national lines with an official source are used; for a country without one, he offers the emergency number and a trusted person.
| Country | Number | Source |
|---|---|---|
| United States | 988 | 988 Suicide & Crisis Lifeline · 988lifeline.org |
| Canada | 988 | 9-8-8 Suicide Crisis Helpline · 988.ca |
| United Kingdom | 116 123 | Samaritans · samaritans.org |
| Ireland | 116 123 | Samaritans · samaritans.org |
| Israel | 1201 | ERAN · eran.org.il |
| Spain | 024 | Línea 024 · sanidad.gob.es |
| France | 3114 | Numéro national de prévention du suicide · 3114.fr |
| Germany | 0800 111 0 111 | TelefonSeelsorge · telefonseelsorge.de |
| Austria | 142 | Telefonseelsorge · telefonseelsorge.at |
| Switzerland | 143 | Die Dargebotene Hand / La Main Tendue · 143.ch |
| Netherlands | 113 | 113 Zelfmoordpreventie · 113.nl |
| Belgium | 1813 | Zelfmoordlijn 1813 · zelfmoord1813.be |
| Australia | 13 11 14 | Lifeline · lifeline.org.au |
| New Zealand | 1737 | Need to talk? 1737 · 1737.org.nz |
What to do in an emergency
For 26 situations, from cardiac arrest, choking and stroke to fire, gas, flood, air-raid alarms and a scam call, James knows the first words to say, when to call the emergency number, and what never to do. First aid is only ever what the emergency dispatcher instructs; he gives no medicines and no doses. Built on:
- European Resuscitation Council Guidelines 2021 · doi.org
- European Resuscitation Council Guidelines 2021: First aid · doi.org
- American Heart Association and American Red Cross — 2024 First Aid Guidelines · cpr.heart.org
- NHS — Sepsis · nhs.uk
- NHS — Falls · nhs.uk
- NHS — Confusion · nhs.uk
- WHO — Abuse of older people · who.int
- Ready.gov — Home fires · ready.gov
- Ready.gov — Floods · ready.gov
- Ready.gov — Earthquakes · ready.gov
- Ready.gov — Power outages · ready.gov
- Israel Home Front Command · oref.org.il
- U.S. Federal Trade Commission — Report fraud · reportfraud.ftc.gov
- Alzheimers.gov — Tips for caregivers and families · alzheimers.gov
Children and learning
When the person is a child, James changes how he talks and how he teaches. The adult questionnaires are switched off entirely. He never assesses a child's mental state, never asks a child to keep anything from the adults raising them, and if a child says something frightening his whole job is to get a real adult involved. With homework he helps the child work it out instead of handing over the answer.
- Asking the child to explain the idea in their own words.Chi MTH, Bassok M, Lewis MW, Reimann P, Glaser R. Self-explanations: how students study and use examples in learning to solve problems. Cogn Sci 1989;13(2):145-82. (experimental study) · DOI
- Small steps, checking understanding after each one.Rosenshine B. Principles of instruction: research-based strategies that all teachers should know. American Educator 2012;36(1):12-19,39. (narrative review) · eric.ed.gov/?id=EJ971753
- Practising by recalling rather than rereading.Karpicke JD, Roediger HL 3rd. The critical importance of retrieval for learning. Science 2008;319(5865):966-8. (randomised experiment) · PubMed 18276894 · DOI
- Coming back to material after a gap instead of cramming.Cepeda NJ, Pashler H, Vul E, Wixted JT, Rohrer D. Distributed practice in verbal recall tasks: a review and quantitative synthesis. Psychol Bull 2006;132(3):354-80. (meta-analysis) · PubMed 16719566 · DOI
- A worked example before working alone.Sweller J, Cooper GA. The use of worked examples as a substitute for problem solving in learning algebra. Cogn Instr 1985;2(1):59-89. (experimental study) · DOI
- Praising the effort and the method, not 'being clever'.Mueller CM, Dweck CS. Praise for intelligence can undermine children's motivation and performance. J Pers Soc Psychol 1998;75(1):33-52. (experimental study) · PubMed 9686450 · DOI Gunderson EA, Gripshover SJ, Romero C, Dweck CS, Goldin-Meadow S, Levine SC. Parent praise to 1- to 3-year-olds predicts children's motivational frameworks 5 years later. Child Dev 2013;84(5):1526-41. (longitudinal study) · PubMed 23397904 · DOI
- Treating a mistake as information, not failure.Dweck CS. The perils and promises of praise. Educational Leadership 2007;65(2):34-9. (expert opinion (review article)) · ascd.org
- Putting the feeling into words.Gottman JM, Katz LF, Hooven C. Parental meta-emotion philosophy and the emotional life of families: theoretical models and preliminary data. J Fam Psychol 1996;10(3):243-68. (longitudinal study) · DOI Lieberman MD, Eisenberger NI, Crockett MJ, Tom SM, Pfeifer JH, Way BM. Putting feelings into words: affect labeling disrupts amygdala activity in response to affective stimuli. Psychol Sci 2007;18(5):421-8. (experimental study) · PubMed 17576282 · DOI
- Waiting a few seconds after a question.Rowe MB. Wait time: slowing down may be a way of speeding up! J Teach Educ 1986;37(1):43-50. (narrative review) · DOI
How we check
- Automated tests on every change. Each condition rule must name its source, and the tests check that every source has a PubMed, DOI or official identifier. Other tests check the safety rules: no doses, the right emergency number for the country, and a real person involved when it matters.
- Scripted conversations with the live model before release. Hard situations, such as a dose question in several languages, words about not wanting to live, a fall or grief, are played to the real voice model.
- An independent judge. Each answer is scored against written criteria by a separate AI model that did not produce the answer.
- References checked by identifier. Every PubMed number and DOI on our pages is checked against PubMed and doi.org before it is published.
Limits
- James is an AI. He says so when asked and never claims to be a person.
- He is not a doctor, nurse, psychologist or therapist, and not an emergency service.
- He can be wrong. He can mishear, misunderstand or miss something. When he is not sure, he asks.
- He does not diagnose, does not advise on doses or treatment, and does not replace the people around you.
- Research describes groups of people, not one person. Some of his conversational rules are our product choices or expert opinion rather than trial results, and we label them so.
- He knows only what the person and the family told him. He never guesses a condition from how someone speaks.
- If a life is in danger, call your country's emergency number.
Rule #1, above everything else. James never helps with medicine doses and never advises treatment. He does not diagnose. He is not a doctor, a therapist or an emergency service: the doctor and the pharmacist decide, and James helps the person reach them.
Set up James