journalism that swims
against the current.
Science + Tech

How to Rebuild Public Health? Don’t Ask the Chatbots

Here’s a case for bottom-up solutions that ignore conventional wisdom.

Crawford Kilian 29 Mar

Tyee contributing editor Crawford Kilian blogs about public health here.

I recently taught a course for ElderCollege at Silver Harbour Seniors’ Activity Centre in North Vancouver called “Post-Pandemic Public Health.” The course title assumed the current pandemic will someday end, but public health needs will not.

Over the first three sessions, I tried to describe the origins of public health, how agencies like the Public Health Agency of Canada had responded to COVID-19, what diseases and social issues could pose future threats, and how we’d have to solve response problems before the next pandemic ambushes us.

The final session was postponed for a couple of weeks, and in the interim I reconsidered my original plan for a 90-minute rant that would give my students the admittedly biased views of one grumpy old man. Perhaps I should offer a broader perspective, and no one has a broader perspective than ChatGPT and Microsoft’s Bing (now souped up with GPT-4).

I went to ChatGPT, and asked: How would you rebuild Canadian public health? Here’s how ChatGPT responded:

As an AI language model, I do not have personal opinions or beliefs, but I can provide general information and suggestions for how to improve public health in Canada based on available data and research.

Here are some ways that Canadian public health could potentially be rebuilt:

ChatGPT then offered five numbered points: increase funding, focus on preventive care, invest in mental health services, strengthen data collection and analysis and address health inequalities. (The bots’ full replies are available as a PDF in the recommended reading section of session four in my course blog.)

Impressed, I went to Bing for a second opinion on rebuilding Canadian public health. Bing replied:

That’s a big question. There are many possible ways to rebuild Canadian public health, but here are some ideas based on recent reports and announcements:

  • Investing more in research to increase public health capacity and innovation.
  • Implementing the recommendations of the Naylor Report, which called for strengthening public health leadership, coordination, surveillance, emergency preparedness and response.
  • Following the vision of the Chief Public Health Officer, which outlined six priorities for transforming public health, such as enhancing equity, resilience, collaboration and accountability.
  • Hiring more doctors, nurses and other health-care workers to reduce wait times and improve access to care.

Now I was more than impressed. Bing mentioned the Naylor report, which was studiously ignored when the Harper government created the Public Health Agency of Canada. Bing also cited our current chief public health officer Dr. Theresa Tam, and even provided footnotes linking to the sources for each point.

Then Bing gave me a teacherly nudge: “What aspect of public health are you most interested in?”

An interrogation had turned into a conversation. I replied: “Maximizing population health, especially for older adults.”

Bing reeled off more bullet points, each footnoted: promote healthy aging, support different living arrangements, encourage “healthy living behaviours” in older people. They could have come out of any number of reports and articles.

Caution and confidence

I went back to ChatGPT with Bing’s question, and got a longer answer, again without footnotes. But it was again familiar advice: promote healthy lifestyles, increase social support, equitable access to health care, and so on.

If the two AIs had a temperamental difference, it was ChatGPT’s caution in phrases like “strategies could be employed,” “may be possible,” “can help.” Bing, more confident, just flat-out advised me to do things: “promote, “implement,” “support.”

But neither had anything new to say. Both offered bromides that could have come from any health ministry’s communications office (or the office of the opposition health critic). Instead they had come from large language models built on reading vast quantities of text about “rebuilding Canadian public health” and “maximizing population health, especially for older adults.” Then it gave me a kind of Reader’s Digest or CliffsNotes version of what it had read.

The chatbots had literally read the room, distilling a consensus that I had heard from experts over and over again since the pandemic began. I suspect my own articles had contributed to that consensus.

No doubt the chatbots had also read the writings of John Kenneth Galbraith, who long ago coined a devastating phrase: “the conventional wisdom.” But they hadn’t read that phrase enough to factor it in.

The conventional wisdom is a solidly argued, even self-evident justification for the status quo. To question the conventional wisdom is to exclude oneself from the serious thinkers of the day, if not to outlaw oneself altogether. The conventional wisdom is groupthink gone amok on a national or even global scale.

And conventional wisdom about rebuilding public health was what the artificially stupid chatbots had given me.

Stop thinking like a doctor?

At about the same time that I was seeking counsel from the chatbots, I ran across a pair of articles by Dr. Eric Reinhart of Harvard University.

One, in the Nation, was titled "Want to Fix Public Health? Stop Thinking Like a Doctor." The other, on the medical website STAT, advised us that “Fixing U.S. Public Health Will Require a Health-Systems Revolution — And for Physicians to Take a Backseat.”

As a political anthropologist, psychoanalyst and medical doctor, Dr. Reinhart has weighed his colleagues in the balance and found them wanting. His unconventional, if not heretical, wisdom is that doctors and most other health-care workers are focused on the sickness of the individual, not the health of the population.

Doctor-think costs lives

“While the clinical frameworks that characterize medical training are appropriate for the one-to-one encounters of patient care,” Reinhart says, “misapplying them to the population-level problems of public health leads to a failure to effectively anticipate and address the social conditions upon which disease and disability feed. This, in turn, fuels a top heavy, reactive national health policy that prioritizes profitable medical treatment rather than cost- and life-saving prevention via community-based social services. Declining life expectancy in the U.S. — now at its lowest in nearly two decades — reflects the consequences of this policy choice.”

Rebuilding public health, Reinhart argues, “will require reclaiming it from its biomedical perversions. In place of physicians who are world-leading experts in narrow doctor-think, the country needs public health systems led by collectives with training across the range of social and biological sciences that effective public health policymaking must synthesize.”

This is public health rebuilt from the bottom up, not from the top down. Among the collectives’ leaders should be those most at risk of disease, disability and shortened lifespans — the poor, the racialized, the unhoused, the imprisoned, and all the other marginalized groups we rarely notice unless we belong to them.

Under Reinhart’s guidelines, experts in the social and biological sciences would sit down with the marginalized to listen and learn. They would then, in close partnership with the marginalized, develop a public health system built on transforming the social determinants of health so that they could move whole populations into more robust health.

Such a system would likely involve womb-to-tomb health care, starting with free prenatal care and continuing with some kind of universal basic income or at least paid parenthood, good education (with free breakfast and lunch), and access to lifelong health care. It would very likely involve a lot more.

Bottom-up messaging

The chatbots’ advice boiled down to messaging: exhorting the public to pay more for public health, exhorting health-care workers to collaborate and be accountable, exhorting elders to display healthy living behaviours, exhorting politicians to address inequalities. The real message was, “everything will be fine if people do as they’re told.”

Reinhart’s bottom-up argument is also for messaging, but coming from the people most at risk of poor health, disabilities and shortened lives. Professionals and rest of the public would have something to say, but it would have to sound like “OK, here’s how we could do it” — not like “please try to be realistic.”

The professionals’ answers would also have to avoid “but that’s socialism!” The health professions have known for over a century that inequality breeds disease; they just haven’t cared, or haven’t cared enough.

Medical apathy toward the poor is an old tradition. After Dr. Rupert Blue stopped bubonic plague in San Francisco, he became the only person to serve as both surgeon general and president of the American Medical Association at the same time. In 1915, he called for a national health insurance program. Public health doctors loved the idea, but those in private practice hated it. Similarly, when Tommy Douglas first brought medicare to Saskatchewan in 1962, the province’s doctors went on strike.

Eric Reinhart’s bottom-up ideas may not be right, but they deserve attention precisely because they’re not conventional wisdom. We have just seen a three-year worldwide demonstration of the gross inability of top-down public health ministries, using conventional wisdom, to protect their citizens’ health. Conventional wisdom has cost at least seven million lives and counting.

So we are overdue for a radical reconsideration of the health of nations. Unless someone tweaks a large language model out of all recognition, chatbots will never come up with a truly public health-care system.

And neither will doctors until they start listening to their patients.  [Tyee]

Read more: Health, Science + Tech

  • Share:

Facts matter. Get The Tyee's in-depth journalism delivered to your inbox for free

Tyee Commenting Guidelines

Comments that violate guidelines risk being deleted, and violations may result in a temporary or permanent user ban. Maintain the spirit of good conversation to stay in the discussion and be patient with moderators. Comments are reviewed regularly but not in real time.


  • Be thoughtful about how your words may affect the communities you are addressing. Language matters
  • Keep comments under 250 words
  • Challenge arguments, not commenters
  • Flag trolls and guideline violations
  • Treat all with respect and curiosity, learn from differences of opinion
  • Verify facts, debunk rumours, point out logical fallacies
  • Add context and background
  • Note typos and reporting blind spots
  • Stay on topic

Do not:

  • Use sexist, classist, racist, homophobic or transphobic language
  • Ridicule, misgender, bully, threaten, name call, troll or wish harm on others or justify violence
  • Personally attack authors, contributors or members of the general public
  • Spread misinformation or perpetuate conspiracies
  • Libel, defame or publish falsehoods
  • Attempt to guess other commenters’ real-life identities
  • Post links without providing context


The Barometer

How Do You Read Your Books?

Take this week's poll