LoopWHI ’26

Health Information Sources

0
9
1
,
2
0
3
9
4
7
5
6
r
7
e
8
s
9
p
10
o
11
n
12
d
13
e
14
n
15
t
16
s
17
18
19
20
6
21
,
22
9
23
9
24
6
25
26
a
27
n
28
s
29
w
30
e
31
r
32
e
33
d
34
35
t
36
h
37
e
38
39
h
40
e
41
a
42
l
43
t
44
h
45
-
46
i
47
n
48
f
49
o
50
r
51
m
52
a
53
t
54
i
55
o
56
n
57
-
58
s
59
o
60
u
61
r
62
c
63
e
64
65
q
66
u
67
e
68
s
69
t
70
i
71
o
72
n

Key takeaways

  • A doctor is still the first place urban working Indians go for health information — 54.0% name one. No other source clears 25%. The clinician hasn't been displaced.
  • AI tools are already the second-ranked source, at 19.8% — narrowly ahead of Google search (19.7%). A category that didn't exist in clinics three years ago now outranks the search engine that defined the last two decades of self-diagnosis.
  • AI use collapses with age: 24.8% of those in their 20s, 7.9% of those in their 50s. It rises just as cleanly with income — 16.4% under ₹3 lakh, 29.9% above ₹40 lakh.
  • 21.6% of workers in their 20s rely on the internet, AI, or influencers without naming a clinician at all — twice the rate of those in their 40s. The youngest cohort is the most likely to skip the doctor entirely.
  • Workplace wellness programs reach 8.0% as a health-information source — the lowest-ranked channel on the list. Employers fund the benefit, then lose the information layer to a chatbot.

A doctor or healthcare professional is the single most-cited source of health information among urban working Indians, named by 54.0% of the 6,996 respondents who answered the question in WHI 2026. That majority sits against a documented Indian backdrop of routine self-diagnosis: a cross-sectional study in central Karnataka found 73% of patients had used online health information to self-diagnose, and 92% reached for Google to do it (Shekhar & Javalkar, IJCMPH, 2020). The clinician leads, but the clinician now shares the room.

Source Share who cite it
Doctor or healthcare professional 54.0%
Doesn't actively seek health information 24.8%
AI tools (ChatGPT, etc.) 19.8%
Google search 19.7%
Family and friends 19.2%
Instagram / YouTube influencers 13.8%
Workplace wellness programs 8.0%
News articles / media 7.4%

The doctor still leads — and AI has already passed Google

Share citing each health-information source, ranked

n = 6,996
Doctor or healthcare professional
54.0%
Doesn't actively seek health information
24.8%
AI tools (ChatGPT, etc.)
19.8%
Google search
19.7%
Family and friends
19.2%
Instagram / YouTube influencers
13.8%
Workplace wellness programs
8.0%
News articles / media
7.4%
Values in %

Multi-select; shares are independent and do not sum. AI tools (19.8%) and Google search (19.7%) are within a tenth of a point — a finding, not a rounding accident. New question for 2026; no year-on-year line.

The question allowed multiple selections, so the columns are independent — a respondent who consults a doctor and also checks ChatGPT is counted in both. The shares describe how broadly each channel reaches into the population, not how a fixed pie is divided.

Two readings of the ranking are worth separating. The first is the headline most observers expect: the doctor still wins. The second is the one underneath it. AI tools, a category that did not exist for consumers before late 2022, are already the second-named source for this workforce — fractionally ahead of Google search, the engine that has organised health-seeking for twenty years. The substitution is not from doctor to machine. It is from search box to chatbot, happening inside a single survey cycle.

AI use is a clean age gradient: 25% in the 20s, 8% in the 50s#

AI tools are not adopted evenly. The share naming ChatGPT or a similar tool falls by roughly a third with each decade of age, from a quarter of those in their 20s to one in twelve of those in their 50s.

Age band Sample Cite AI tools Cite a doctor
20–29 2,830 24.8% 50.7%
30–39 2,977 18.3% 55.4%
40–49 923 11.8% 59.6%
50–59 216 7.9% 58.8%

AI fades with age while the doctor strengthens

Share citing AI tools vs a doctor, by age band

Cite a doctorCite AI tools

A worker in their 20s is three times more likely to cite AI than a worker in their 50s (24.8% vs 7.9%) — and the least likely of any band to cite a doctor (50.7%). 50–59 cell n = 216.

The two columns move in opposite directions. AI reliance falls with age while doctor reliance climbs — the youngest cohort is both the heaviest user of AI and the lightest user of clinicians. The gap is wide: a worker in their 20s is three times more likely to cite AI than a worker in their 50s, and meaningfully less likely to cite a doctor.

This is the generational fault line the rest of the page sits on. It is not a story about one app. It is a story about which authority a person reaches for first when something feels wrong — and that default is being set, for the cohort now entering the workforce, by a tool that answers instantly and never asks them to book an appointment. It is also the channel through which this workforce is meeting new therapies: the GLP-1 Awareness page finds awareness of those drugs concentrating in the same young, high-income, digitally-fluent cohort that leans hardest on AI here.

AI reliance rises with income; doctor reliance rises with it too#

The same gradient runs along income, and in the same direction for AI.

Income band Sample Cite AI tools Cite a doctor
Under ₹3 lakh 876 16.4% 46.5%
₹3–6 lakh 1,981 18.0% 49.4%
₹6–10 lakh 1,401 18.8% 53.5%
₹10–15 lakh 958 21.1% 56.9%
₹15–25 lakh 732 25.7% 62.2%
₹25–40 lakh 373 24.7% 63.0%
₹40 lakh+ 137 29.9% 69.3%

Higher income buys more of both channels

Share citing AI tools and a doctor, by income band low to high

Cite AI toolsCite a doctor

AI is not a substitute forced by cost — the highest earners use the most of both (29.9% AI, 69.3% doctor; n = 137). The lowest band consults fewer sources of any kind.

The intuitive prior — that AI is what people reach for when they can't afford a doctor — does not hold in this data. Both rise together with income. The highest earners are simultaneously the most likely to consult a clinician (69.3%) and the most likely to consult an AI (29.9%). AI is not, for this workforce, a substitute forced by cost. It is a supplement adopted by the same people who already have the most access to care.

The bottom income band tells the inverse story and is the one worth holding onto. Workers earning under ₹3 lakh are the least likely to name a doctor (46.5%) and among the least likely to name AI (16.4%). They aren't trading one source for another. They are simply consulting fewer sources of any kind — which is consistent with the 24.8% across the whole sample who say they don't actively seek health information at all.

One in four doesn't go looking at all#

The second-largest answer on the page is an absence. 24.8% of respondents say they don't actively seek health information from any source. That is a larger group than the one citing AI, Google, family, or influencers individually.

Disengagement by age Sample "Don't actively seek"
20–29 2,830 27.0%
30–39 2,977 24.0%
40–49 923 21.7%
50–59 216 20.4%

Disengagement peaks in the 20s

Share who say they don’t actively seek health information, by age band — line marks the sample average

20–29
n = 2,830
27.0%
30–39
n = 2,977
24.0%
40–49
n = 923
21.7%
50–59
n = 216
20.4%

Sample average — 24.8%

Values in %

The same cohort that uses AI most heavily is also the most likely to seek nothing — a young workforce split between digital self-direction and switched-off.

Disengagement is highest in the 20s and falls steadily with age — the same cohort that uses AI most heavily is also the one most likely to seek nothing. The two facts are not in tension. They describe a young workforce that splits between active digital self-direction and passive non-engagement, with the steady, doctor-anchored middle thicker among older workers. Health-seeking, in the 20s, is either intense and self-directed or simply switched off.

The youngest cohort is the most likely to skip the clinician entirely#

The multi-select format lets us ask a sharper question than "who uses AI". It lets us ask who uses the internet, AI, or an influencer for health information without naming a clinician at all — a self-directed digital channel with no professional in the loop.

Age band Sample Digital sources, no clinician named
20–29 2,830 21.6%
30–39 2,977 16.6%
40–49 923 10.6%
50–59 216 11.6%

A fifth of twenty-somethings navigate with no clinician at all

Share citing a digital source (search, AI, or influencer) with no doctor anywhere in their mix, by age band

20–29
n = 2,830
21.6%
30–39
n = 2,977
16.6%
40–49
n = 923
10.6%
50–59
n = 216
11.6%
Values in %

Digital source cited, doctor not cited. The share halves by the 40s — the cleaner read of the generational shift than the raw AI number.

More than one in five workers in their 20s sources health information from a screen — search, AI, or an influencer — with no doctor anywhere in their answer. That share halves by the 40s. This is the cleaner read of the generational shift than the raw AI number: it isn't just that young workers add AI to their mix, it's that a fifth of them have a mix with no clinician in it.

Influencers reach one in seven — and skew female and young#

Instagram and YouTube health influencers are the sixth-ranked source overall at 13.8%, but the headline understates the concentration. The channel is meaningfully more female and more young.

Sample Cite influencers
Women, 20–29 1,045 17.0%
Men, 20–29 1,776 15.7%
Women, 40–49 250 13.2%
Men, 40–49 672 7.6%

Influencer reach holds for women and halves for men

Share citing Instagram / YouTube influencers, by gender and age band

WomenMen

Roughly gender-balanced in the 20s (17.0% vs 15.7%), distinctly female by the 40s (13.2% vs 7.6%). Women 40–49 n = 250 · men 40–49 n = 672.

Among men, influencer reliance more than halves between the 20s and the 40s, from 15.7% to 7.6%. Among women it barely moves. The result is a channel that is roughly gender-balanced for the youngest workers and distinctly female by mid-career. Women also out-cite men on Google search (23.1% vs 18.1%) and on doctors (59.5% vs 51.6%) — across the board, women in this workforce report consulting more health-information sources than men do.

Workplace wellness is the weakest channel — 8.0%, last on the list#

The source employers most directly control is the one workers least often name. Workplace wellness programs register as a health-information source for 8.0% of respondents — the bottom of the ranking, below influencers, below family and friends, a quarter of the doctor's reach and well under half of AI's.

The gap is the finding. Employers fund the health benefit, the screening, the insurance. Then the information layer — what an employee actually does with a symptom, a result, a worry at 11pm — routes to a chatbot, a search bar, or an influencer the employer has no relationship with. The wellness channel reaches its high-water mark of 12.2% in Pharma and bottoms out at 4.3% in Logistics; in no industry does it approach the AI number for the same workforce.

The clinician leads, but the default is being reset#

The doctor is still first. That is the honest headline and it should not be buried: 54% of urban working Indians name a clinician, and no other source comes close. But the page underneath the headline shows where the next decade is heading. AI has already passed Google for the second slot. The youngest, highest-earning workers are adopting it fastest, and a fifth of those in their 20s are navigating their health from a screen with no professional named at all.

References

  1. 1Ayre, J., Cvejic, E., McCaffery, K. et al. Use of ChatGPT to obtain health information in Australia, 2024: insights from a nationally representative survey. Medical Journal of Australia 222, 175–177 (2025). https://doi.org/10.5694/mja2.52598
  2. 2Shekhar, R. & Javalkar, S.R. Health information seeking on the internet (Dr Google) and its effect on doctor-patient relationship: a cross-sectional study from Central Karnataka. International Journal of Community Medicine and Public Health 7, 4076–4080 (2020). https://doi.org/10.18203/2394-6040.ijcmph20204379
  3. 3Shool, S., Adimi, S., Saboori Amleshi, R. et al. A systematic review of large language model (LLM) evaluations in clinical medicine. BMC Medical Informatics and Decision Making 25, 117 (2025). https://doi.org/10.1186/s12911-025-02954-4
  4. 4Bhattacharyya, M., Miller, V.M., Bhattacharyya, D. & Miller, L.E. High rates of fabricated and inaccurate references in ChatGPT-generated medical content. Cureus 15, e39238 (2023). https://doi.org/10.7759/cureus.39238