In a new survey, 60% of US adults say they would feel uncomfortable if their own doctor used AI for things like diagnoses and treatments, against 30% who would feel comfortable, according to a Pew Research Center report published on 29 October. The survey of 3,488 adults ran from 22 to 28 June 2026.
The sharpest move is among the young. Pew says 45% of adults under 30 now expect AI to lead to worse patient outcomes, up from 29% in 2022, while the share expecting better outcomes fell from 44% to 31%. The group’s net view swung by 29 points, from 15 points positive to 14 points negative, by our arithmetic.
Business Pill · Guardrails
An evergreen Business Engineer explainer on guardrails. It teaches the general idea only and does not discuss the Pew survey.
Adults aged 18 to 29 who say using AI for diagnoses and treatments will lead to better or worse patient outcomes, 2022 and 2026. Data: Pew Research Center, report published 29 October 2026 (2026 survey of 3,488 US adults, 22 to 28 June). Chart: FourWeekMBA.
The key insight: Pew finds 60% of US adults would be uncomfortable with their own doctor using AI for diagnoses and treatments, and only 3% are open to AI making a diagnosis on its own; adults under 30 swung from 15 points positive to 14 points negative on AI’s effect on patient outcomes since 2022, a 29-point swing by our arithmetic.
Pew writes that Americans still “want a human in the loop”. Openness depends on the task. About eight in ten (78%) are open to AI translating for a patient, either with a human final review (52%) or on its own (26%).
At the other end are diagnoses and insurance claims. Just 3% are open to AI making a diagnosis on its own, while 47% would accept it with a human final review and 41% are not open at all. For insurance claim decisions, “nearly half (46%) do not want AI to have anything to do with making insurance claim decisions”, and 5% are open to AI deciding on its own, about nine times fewer, by our arithmetic.
Share of US adults aged 18 to 29 who say using AI for things like diagnoses and treatments will lead to worse patient outcomes: 29% in 2022 and 45% in 2026, from Pew Research Center surveys (2026 survey of 3,488 adults, 22 to 28 June, published 29 October 2026). The 16-point change is our arithmetic.
What Was Posted on X
New study from PEW Research
Majority of Americans uncomfortable with their doctor using AI
> Americans want humans making the final call: 47% accept AI diagnoses with human review, but only 3% would let AI diagnose alone, and 46% reject any AI role in insurance claims.
> 60% would feel uncomfortable if their own doctor used AI for diagnoses and treatments, double the 30% comfortable, even though 78% accept AI translating for patients.
> Americans split evenly on AI’s effect on patient outcomes, 37% better versus 36% worse, but 45% of adults under 30 now expect worse outcomes, up from 29% in 2022.
> Privacy and the patient-doctor relationship draw the most pessimism: 51% expect AI to worsen health-data security against 8% better, and 47% expect a worse relationship against 10% better.
> Trust in oversight is thin: 58% doubt health insurers and 50% doubt AI companies will keep AI serving patients, while only 25% doubt doctors.
Rohan Paul’s post on X of 4 November relaying the Pew Research Center study. It says 60% would feel uncomfortable if their own doctor used AI for diagnoses and treatments, and that 45% of adults under 30 now expect worse outcomes, up from 29% in 2022. The piece checks every figure against Pew’s own report.
Split on Outcomes, Pessimistic on Privacy
Overall, 37% say AI in diagnoses and treatments will lead to better patient outcomes and 36% say worse, with 26% expecting little difference. Pew says these views have “largely remained unchanged since we last asked this in 2022”; Pew also reports that some groups have moved. Adults 65 and older moved the other way: 42% now expect better outcomes, up 8 points.
On data and the doctor’s office, the balance is negative. 51% say AI will make the security of patients’ health information worse, against 8% better, and 47% expect a worse patient-doctor relationship, against 10% better. More say AI will improve access to services than worsen it, 31% to 20%.
Who Is Trusted to Keep AI on the Patient’s Side
Insurers rank last. Pew finds 58% have not too much or no confidence in health insurance companies to make sure AI is used in patients’ best interest, followed by the federal government (55%), AI companies (50%) and state governments (50%). Doctors fare best: 25% have low or no confidence in them, and 26% a great deal or quite a bit.
Insurance claims are also the task where the survey finds the strongest objection to any AI role. On the same day this study was relayed on X, CVS Health said it plans to split in two, separating its Aetna insurance business from CVS Pharmacy.
The figures were relayed on X on 4 November by Rohan Paul, whose post summarised the report’s main numbers; this piece uses Pew’s own report and methodology for every figure.
As we read the report, acceptance tracks the stakes: 78% are open to AI translating for a patient, while 46% want no AI role in insurance claim decisions.
As we read the confidence question, insurers rank lowest: 58% have little or no confidence in them, followed by the federal government (55%), with AI companies at 50%, against 25% for doctors.
Pew Research Center, 29 October 2026
“Americans still want a human in the loop”
Three Implications
FOR HEALTH INSURERS Insurance claim decisions draw the strongest objection: 46% do not want AI to have anything to do with them, and 58% have little or no confidence in insurers on AI.
FOR AI VENDORS Half (50%) have little or no confidence in AI companies to make sure AI is used in patients’ best interest.
FOR DOCTORS Doctors are the most trusted actor in the survey; human review draws the largest shares for tasks such as translating (52%), analysing medical scans (60%) and making a diagnosis (47%).
The framework’s FRED Test scores a workflow on frequency, repeatability, error tolerance and decomposability, and says: “Score high on all four and the workflow is automated within the cycle; score low on any one and the human stays in the loop.”
As we read Pew’s numbers, the public draws a similar line: the tasks with the least room for error, diagnoses and insurance claims, get the least support for AI acting alone. The framework does not discuss this survey.
The survey measures attitudes, not behaviour: it does not show whether patients would avoid doctors or insurers that use AI. Pew’s methodology gives a margin of sampling error of plus or minus 1.8 percentage points for the full sample of 3,488. The report was published on 29 October and was relayed on X on 4 November. The 29-point swing and the nine-times ratio are our arithmetic on Pew’s figures.
This Week In The AI Supercycle
Anthropic’s Reality Check, The Capital Squeeze, & The New World Order
This week’s issue: for two years the AI build-out was priced as a growthstory. This week it started being priced as a credit story.
Pew finds 60% of US adults would be uncomfortable with their own doctor using AI for diagnoses and treatments, 46% want no AI role in insurance claim decisions, and 58% have little or no confidence in insurers to keep AI working for patients. Adults under 30, more optimistic than others in 2022, now lean negative: 45% expect worse outcomes, a 29-point net swing by our arithmetic.
A note on sourcing. All survey figures and quotations are from Pew Research Center’s report and methodology published on 29 October 2026 (survey of 3,488 US adults, 22-28 June 2026). Point changes, the net swing and ratios are our arithmetic.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
