Can pharma stay relevant in the age of AI?
Four years after its launch, OpenAI's ChatGPT now reaches more than a billion people a month. The way the world looks for information has changed for good, and medicine is no exception: alongside the journal and the search bar, doctors are increasingly turning to the chat window.
In this report, we follow that shift: where HCPs turn for answers today, how patients are changing the conversation and where pharmaceutical companies fit in an increasingly AI-driven world.
Research methods
EMJ GOLD conducted a targeted online survey distributed via email to healthcare professionals, receiving 60 responses from participants representing a range of ages, medical specialities, genders and nationalities.
Most trusted sources by HCPs
When a clinical question arises, most healthcare professionals have a go-to source, and for many, it's a familiar one. Asked which established mediums they turn to first, respondents put the medical journal at the top of the list, while pharma resources came in last.
Only 2% turn to pharma first
Search engine
Medical journal
Decision-support tool
Colleague or peer
Online community
Pharma resource
Guideline body
% of respondents; multiple answers allowed
Source: EMJ, 2026
This data points to a significant challenge for the industry. Pharma resources may be in the mix, but they are rarely the first place a healthcare professional looks. Instead, HCPs start with independent sources: journals, decision-support tools and guideline bodies. To understand why, it helps to look at what HCPs weigh up when selecting a source.
Trust wins out over speed and convenience
What drives resource choice
Which three factors most drive which source you reach for in the moment?
%
%
% of respondents choosing each factor; up to three answers allowed
Source: EMJ, 2026
%
Clinical depth
Trustworthiness
Speed or convenience
Trust is the clear deciding factor, cited by 76% of respondents, with clinical depth
a distant second at 48%. For clinicians, where an answer comes from matters as much as the answer itself.
Speed and convenience rank third, chosen by 44%. Even though they are time poor, HCPs appear to put trustworthiness above all else. On paper, that should be a hurdle for AI tools, whose main appeal is speed. In practice, it hasn't stopped clinicians from using them.
Adoption of AI tools
AI is already part of everyday practice
85%
of HCPs have used an AI tool for clinical or professional work in the last three months
33%
AI answers in a search engine
26%
Medical-specific AI tools
9%
AI in a decision- support platform
4%
Ambient scribe or documentation AI
70%
General AI assistant
% of respondents; multiple answers allowed
Source: EMJ, 2026
39%
31%
30%
30%
22%
4%
2%
AI use is already widespread, but it's led by everyday tools rather than specialist ones. Almost three times as many respondents have used a general assistant as a medical-specific tool, with some turning to AI much as they would to a colleague. Tools built into clinical systems, by contrast, have yet to take hold: just 9% have used AI within a decision-support platform, and 4% an ambient scribe. For now, AI adoption appears to be led by individual clinicians rather than their institutions. But using a tool is not the same as trusting it.
"My second brain."
— Respondent, 29 years in practice
"AI is already part of my clinical practice, helping me rapidly access and synthesise evidence and supporting, but never replacing, my clinical judgment." — Respondent, 50 years in practice
usually or always verify AI-generated answers
8 in 10
Clinicians use AI, but with a key caveat
Don't trust or avoid
46%
33%
11%
9%
Cautious, always verify
Trust, usually verify
Trust, rarely verify
Trust in AI-generated answers
"If verified, it is very helpful."
— Respondent, 18 years in practice
Despite the widespread adoption, caution is evident in how clinicians use AI. That raises a question about efficiency: if most answers have to be checked against another source, much of the time AI saves may be spent again on verification.
And the verification burden doesn't stop with clinicians' own searches. Increasingly, AI is shaping what patients bring into consultations too.
Patients arrive with answers of their own
Patient use of AI
How often do patients arrive with information from AI sources?
54%
2%
30%
15%
Frequently, and increasing
Rarely
Never
Occasionally
In the survey, more than half of respondents said patients frequently arrive with information from AI tools or online sources, and that it's happening more often. Only 2% said it never happens. When asked whether this was a good thing, the jury was out on its benefits.
Clinicians are split on patients using AI
28%
Mixed – helps and hinders
28%
46%
19%
Mostly positive –
better-informed patients
say patient use of AI mostly creates misinformation they then have to correct
Mostly negative –
creates misinformation to correct
While HCPs see both benefits and drawbacks to AI, concerns lean negative. They must not only scrutinise what AI tells them, but also help patients separate fact from fiction.
"[AI] creates anxiety for patients who 'self-diagnose'."
— Respondent, 35 years in practice
A gap pharma could fill
The need for verification, coupled with concerns about misinformation, is where pharma companies could play a valuable role. HCPs need reliable sources they can access quickly and easily, with confidence in the evidence they provide. But that role depends on something pharma has yet to earn: HCPs' trust.
Pharma's trust gap is real, but repairable
81%
THE GAP
THE BRIDGE
say knowing content is funded or produced by pharma reduces their trust
What would most increase your trust in pharma-produced information?
67%
52%
26%
20%
Full data transparency, including unfavourable data
Peer-reviewed publication
Co-created with independent experts or societies
Upfront funding disclosure
Source: EMJ, 2026
% of respondents; single answer
Source: EMJ, 2026
% of respondents; single answer
Source: EMJ, 2026
% of respondents; single answer
Source: EMJ, 2026
% of respondents; single answer
That gap is the industry's biggest hurdle. Four in five HCPs say knowing content is funded or produced by pharma reduces their trust in it. Most, however, say it reduces their trust slightly rather than significantly, which suggests the damage can be repaired.
Disclosure alone isn't enough
Declaring who paid for content is not enough on its own, as HCPs want to be able to judge the evidence for themselves.
Could pharma earn a role in AI?
If pharma can meet those conditions, would HCPs use an AI tool built on them?
Openness to a pharma AI tool
Two in three would use a pharma AI tool, on the right terms
19%
48%
19%
6%
4%
What earns trust?
Full data transparency, including unfavourable data, leads by a clear margin, followed by peer-reviewed publication.
Very likely
Likely
Unsure
Very unlikely
Source: EMJ, 2026
% of respondents; single answer
Key takeaway
Two in three HCPs say they would be likely to use a pharma AI tool, a stark contrast with the 2% who currently turn to pharma first with a clinical question. But that openness comes with conditions, and they may not be easy for pharma to meet.
The tool HCPs responded to would draw on fully sourced, independently reviewed evidence, and trust, they say, depends on full transparency, including unfavourable data. That means giving up some control over the message, a step not every company will be comfortable taking. For clinicians, though, the priority is simple.
"I try to embrace any tool that helps me provide better care to my patients."
— Respondent, 45 years in practice
That focus on patient care is the common ground. AI isn't going away; it's only becoming more capable and more widely used, and clinicians are already relying on it. For pharma, the opportunity isn't to win them back from AI, but to earn their trust, either by working with credible partners who can bridge that gap, or by joining the revolution and becoming a trusted AI source they don't feel the need to double check.
Unlikely
Acknowledgements
The following individuals are acknowledged for their valuable contributions to this report:
Isabel O’Brien, Content Manager, EMJ – Research, Writing and Survey
Helena Spicer, Junior Designer, EMJ – Design
Stacey White, Creative Group Head, EMJ – Creative Support
Bora Erdemli, Partner, ZS – Industry Insight
76
48
44