A critical response to “How GLP-1s and GenAI Can Solve Healthcare’s Affordability Crisis”
A new LinkedIn article by Robert Pearl argues that GLP-1 medicines and generative artificial intelligence can solve America’s healthcare affordability crisis. The argument is appealing.
It is also far too confident.
The article correctly identifies chronic disease as a major cause of illness, disability, and healthcare spending. It also correctly recognizes that GLP-1 drugs can produce substantial weight loss and reduce important complications in selected patients, and that artificial intelligence may eventually improve monitoring and continuity of care. These are serious advances. They deserve serious use.
But treatment is not the same as prevention. Clinical efficacy is not the same as affordability. A promising technology is not a health policy. And two expensive, commercially supplied technologies do not constitute a national prevention strategy.
The article makes four unsupported leaps
First, it calls pharmacologic treatment of obesity “prevention.” Treating obesity can prevent diabetes, cardiovascular events, heart-failure complications, and other downstream outcomes. That is valuable secondary prevention. But it does not prevent the social and commercial conditions that help produce obesity in the first place.
Second, it moves from weight loss and disease outcomes to healthcare savings. Those are different questions. A drug can be clinically effective and still increase total spending, especially when it is costly, used by millions of people, and continued for years. The cited trials were designed primarily to test clinical outcomes, not whether nationwide GLP-1 coverage would reduce net healthcare expenditures. After GLP-1 withdrawal, substantial weight regain is common, which further weakens any assumption that a short course will generate durable savings.
Third, it treats early studies of generative AI as evidence that AI can safely manage chronic disease at scale. A study showing that an AI system performs well on written cases or produces recommendations judged favorably is not evidence that an autonomous monitoring system will reduce myocardial infarction, stroke, kidney failure, hospitalization, mortality, or total costs. The randomized diagnostic-reasoning trial cited in the article found no significant improvement in physicians’ scores when they were given access to GPT-4, although the model alone scored higher. That finding is interesting, but it does not validate the proposed national care model.
Fourth, the claim that medicine lacked tools to address chronic disease five years ago is plainly wrong. Long before the current GLP-1 era, we had tobacco-control laws, blood-pressure treatment, statins, vaccines, cancer screening, nutritional policy, exercise interventions, and the Diabetes Prevention Program. The problem was never simply that medicine lacked another product. The problem was failure to implement effective prevention consistently, equitably, and politically.
Real prevention starts before the prescription
America often calls the treatment of a preventable disease “prevention” once it becomes profitable and billable.
That reverses the proper order.
Real prevention means fewer children becoming addicted to nicotine. It means food that supports health is affordable, understandable, and available. It means schools teach nutrition and provide healthy meals. It means communities are built so that walking, cycling, and exercise are safe and ordinary. It means clean air, adequate housing, paid time to care for one’s health, and primary care that begins before metabolic disease is established.
Lifestyle change is not a slogan and should not become a way to blame patients.
Telling someone to “eat better and exercise” while surrounding that person with cheap ultraprocessed food, targeted advertising, unsafe streets, long working hours, and unaffordable fresh food is not prevention.
It is abandonment disguised as advice.
Education is necessary, but education alone is weak policy. People need clear food labeling and sound health education. They also need an environment in which the healthier choice is available, affordable, and practical. The landmark Diabetes Prevention Program showed that an intensive lifestyle intervention reduced progression to diabetes among high-risk adults. Yet intensive programs require time, coaching, access, and sustained support. A pamphlet is not an intervention.
Tobacco policy should be far more aggressive
Cigarettes are not ordinary consumer goods. Used as intended, they cause addiction, disease, and premature death. Governments should move toward ending their commercial sale, while avoiding criminal penalties for people who smoke. The target should be the product and the industry, not the addicted patient.
That means sharply higher excise taxes, comprehensive smoke-free laws, plain packaging and graphic warnings, bans on advertising and sponsorship, strict retailer licensing, enforcement against sales to minors, fully funded cessation treatment, and serious consideration of a phased tobacco endgame. No one should be forced to inhale smoke at work, at home, or in public space.
Alcohol requires stronger regulation, not careless prohibition
Alcohol causes cancer, liver disease, injury, violence, fetal harm, and other preventable morbidity. Its health burden is routinely minimized because drinking is socially accepted and commercially promoted.
Governments should tax alcohol according to ethanol content, restrict hours and density of sales, prohibit or tightly limit advertising and sponsorship, require prominent health warnings, enforce minimum purchase ages, strengthen impaired-driving laws, and protect children from marketing. These are among the evidence-based measures recommended by the World Health Organization.
A blanket ban on alcohol consumption, however, is not supported by the same evidence and could create illicit markets, discriminatory enforcement, and additional harm. Public health should be bold, but it should also be empirical. The defensible aim is to reduce alcohol consumption and exposure through strong regulation, not to repeat the failures of criminal prohibition.
Do not medicalize our way out of a political failure
GLP-1 medicines should be available to patients who meet evidence-based criteria. Denying effective treatment because society failed at prevention would be unethical. Generative AI should also be tested rigorously and used when it demonstrably improves outcomes, access, safety, or workload. Neither position requires technological skepticism.
What requires skepticism is the claim that broader access to drugs and AI will, by itself, make American healthcare affordable. That conclusion has not been demonstrated. It also directs attention toward reimbursable products and away from policies that threaten the profits of tobacco, alcohol, and unhealthy-food industries.
The central affordability question is not merely how to treat more chronic disease at a lower unit cost. It is why the United States tolerates commercial and social conditions that generate so much preventable disease, then pays extraordinary prices to manage the consequences.
We need medications. We need better clinical technology. But prevention is not another prescription. Prevention is education backed by access, healthier food systems, daily movement designed into communities, freedom from tobacco smoke, strong alcohol regulation, and public policy willing to place health ahead of commercial convenience.
If we continue to call downstream treatment “prevention,” we may improve many lives. But we will not have solved the cause, and we should not claim that we have solved the cost.
References
1. Pearl R. How GLP-1s and GenAI can solve healthcare’s affordability crisis. LinkedIn. September 7, 2026.
2. Aronne LJ, Horn DB, le Roux CW, et al. Tirzepatide as compared with semaglutide for the treatment of obesity. N Engl J Med. 2025;393:26-36. doi:10.1056/NEJMoa2416394.
3. Jastreboff AM, le Roux CW, Stefanski A, et al. Tirzepatide for obesity treatment and diabetes prevention. N Engl J Med. 2025;392:958-971. doi:10.1056/NEJMoa2410819.
4. Wilding JPH, Batterham RL, Davies M, et al. Weight regain and cardiometabolic effects after withdrawal of semaglutide: the STEP 1 trial extension. Diabetes Obes Metab. 2022;24:1553-1564. doi:10.1111/dom.14725.
5. Goh E, Gallo R, Hom J, et al. Large language model influence on diagnostic reasoning: a randomized clinical trial. JAMA Netw Open. 2024;7:e2440969. doi:10.1001/jamanetworkopen.2024.40969.
6. Knowler WC, Barrett-Connor E, Fowler SE, et al. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med. 2002;346:393-403. doi:10.1056/NEJMoa012512.
7. World Health Organization. Tackling NCDs: best buys and other recommended interventions for the prevention and control of noncommunicable diseases. 2nd ed. Geneva: World Health Organization; 2024.
8. World Health Organization Regional Office for Europe. Commercial determinants of noncommunicable diseases in the WHO European Region. Copenhagen: WHO Regional Office for Europe; 2024.