In the fall of 2022, a remarkable innovation burst onto the scene with the power to reshape the world and kick-start the next Great Productivity Boom.
That would be ChatGPT, obviously. In November that year, the artificial intelligence chatbot made its debut, attracting a billion visits within a matter of weeks and launching at least as many debates about the future of work, life and humanity itself under AI.
Less obvious was the other technological breakthrough that grabbed the world’s attention at the same time and started literally reshaping its users: Ozempic, a drug developed by Denmark’s Novo Nordisk A/S and approved several years earlier to treat Type 2 diabetes. Ozempic also has the convenient side effect of inducing rapid weight loss in people. Hollywood celebrities twigged to its slimming powers, and the TikTok crowd soon followed.
In the less than two years since AI and Ozempic seemed to erupt out of nowhere, AI’s impact on society and the economy has been endlessly analyzed, gamed out and hyped, and its transformative potential is almost universally accepted.
Ozempic? Not so much. As a society we’re not accustomed to thinking about medicines in the same terms of social and economic change, especially ones that help tackle a problem with as many moral and cultural hangups as chronic obesity. Ozempic, which has come to be a catch-all name for a whole category of GLP-1 type weight-loss drugs, has been variously framed as a medical advancement, diet fad, a symbol of thin-obsessed celebrity culture and an overpriced weight-loss cheat. Anything but an economic disruptor.
That’s now changing as some economists and investment firms come to appreciate Ozempic-like drugs for what they could do to jolt a plodding economy, especially with an explosion of similar drugs in the pipeline and the promise of health benefits that go well beyond weight loss, including lowered risks of heart attacks, kidney disease, cognitive decline and even addiction.
And while Ozempic and AI may seem like an unlikely comparison, contrasting the two breakthroughs for their disruptive capacity and economic impact, as well as the steep barriers to mass adoption each must overcome, sheds light on how unique and historic this moment could be.
Obesity is rising rapidly around the world. More than 40 per cent of Americans are obese, while roughly 30 per cent of Canadians fall into that weight category, and roughly one in three people in both countries are overweight (one level down from obesity). The costs are staggering. In the United States, people diagnosed with diabetes account for one in four health care dollars spent, according to the National Institutes of Health. Studies show obesity is linked to higher unemployment, greater absenteeism from work and, particularly for women, wage discrimination.
When taken, currently as a once-a-week injection, drugs such as Ozempic, as well as Wegovy, Mounjaro and Zepbound, essentially hack the brain, tricking the body into feeling full earlier in meals, and leading users to consume less food and slim down. The average weight loss for people taking such medications is 15 per cent to 20 per cent.
Add it all up and experts are beginning to see the potential for dramatic economic gains from anti-obesity medications, coming on top of an expected productivity boost from AI.
“Normally supply-side shocks to the economy tend to be gradual and drawn out, but with both obesity drugs and AI, you’re seeing these massive positive supply shocks unfold in real time,” said Lars Christensen, a Danish economist who has studied the economic impact of obesity, co-founded an AI consultancy and has seen the direct impact of Ozempic on his country, where Novo Nordisk’s market valuation is now greater than Denmark’s gross domestic product. “We’re at the start of an era where AI and weight-loss medicines have enormous potential to boost long-term productivity growth.”
In other words, AI could drive growth by replacing some humans, while Ozempic could do the same by improving them.
Mass disruption
Anyone who has conversed with ChatGPT knows how other-worldly the experience can be. The app’s approachable interface, powers of reasoning, flexibility and speed have made it a popular tool to do things such as draft and edit reports, write code and conduct research.
You could, for instance, ask ChatGPT to estimate how much weight Americans have collectively shed thanks to Ozempic-like medications, and after factoring in estimates for the number of users and average weight loss rates, it suggests a range of between 15,000 tons and 90,000 tons.
But large language models such as ChatGPT also sometimes fabricate responses. For instance, ask ChatGPT to put that 90,000 tons figure in context, and it will authoritatively tell you this is nearly the weight of an Oasis-class cruise ship. A quick check of Royal Caribbean International’s website shows Oasis ships weigh nearly 227,000 tons, a mistake the AI app readily admits after its error is pointed out.
Yet it would be wrong to use such flubs to dismiss the changes coming with AI.
“Literally every sector of the economy and society are going to be totally transformed by AI and nothing will remain unchanged,” said Rob Toews, a San Francisco-based partner with Radical Ventures, a Toronto venture capital firm that focuses on the sector. “Every board of every Fortune 500 company is now getting very serious about their AI strategy.”
He rhymes off industries and sectors that AI is set to upend, such as software engineering, customer support work and banking. Even more upheaval is on the horizon with innovations around autonomous agents: programs capable of tackling high-level tasks without being prompted and prodded like ChatGPT must be.
“Pretty soon we’re going to see AI products that are not like tools to help humans increase their productivity, but AI agents that do everything a human would do for a given job,” he said.
Interestingly, health care is one field in which Mr. Toews sees some of the biggest opportunities, since AI excels at sifting through massive troves of data and spotting patterns humans simply can’t. For instance, in 2020, Google DeepMind, the tech giant’s AI lab, unleashed its AlphaFold AI model on a problem that had confounded researchers for decades: how to predict and model the three-dimensional shape of individual proteins. The answer was achieved in hours and AlphaFold has pushed molecular science even further since.
For that reason, Mr. Toews bristles at the notion that AI and Ozempic can be compared. The drug is a marvel, he said, but AI will discover the next Ozempic. “Biology will be completely transformed by AI and the ability to make super precise drugs that are personalized for each individual based on their entire genome,” he said.
Yet those developments may still take years to become reality. On the other hand, the science behind Ozempic, and its impact on businesses and on people’s lives, is far more straightforward and immediate.
“For 40 years people have been saying AI has great potential and it always will, but the impacts for now are more hidden and we haven’t seen them in a business or personal sense yet,” said Jeffrey Cole, director of the Center for the Digital Future at University of Southern California Annenberg and a researcher who studies disruptive technologies. “But with Ozempic you see the impact right now – America’s favourite pastime is watching television and guessing who is on these drugs.”
While Ozempic, the brand name for semaglutide, was approved as treatment for Type 2 diabetics, off-label use of it for weight loss exploded in 2023, prompting fears of shortages and rationing for people with diabetes. Ozempic’s counterpart, Wegovy, also developed by Novo Nordisk, was approved specifically for weight loss in Canada in 2021 and hit pharmacy shelves earlier this year after years of supply constraints.
Meanwhile, U.S. pharma giant Eli Lilly and Co. has two similar drugs on the market with the active ingredient tirzepatide. Mounjaro is aimed at diabetics and Zepbound (which is not available in Canada) targets weight loss.
When people eat, their guts release GLP-1 hormones into the bloodstream. That curbs their appetite and stimulates the release of insulin. Semaglutide mimics the GLP-1 hormone while tirzepatide mimics GLP-1 as well as another intestinal hormone, GIP. Doing so makes people feel full much earlier in a meal than they normally would.
Hartley Macklin, who runs a Facebook group for Canadian semaglutide users, noticed the change right away when he began using Saxenda, a daily injection that was a precursor to Ozempic – he’s been on the latter since 2018. “We used to go out for Sunday morning breakfast at a greasy spoon and we soon noticed we could not finish our breakfast like we could before, so we’d share a plate,” he said.
From a peak weight of 300 pounds, the 62-year-old Winnipeger has shed more than 60 pounds.
Take results like that and multiply them by the tens or even hundreds of millions of potential users of GLP-1 drugs, and the scale of disruption comes into focus, said Mr. Cole.
He estimates there are roughly 230 million people who are either overweight or obese in the U.S. who the drugs could help at some point.
The drugs aren’t a silver bullet to shedding weight, and work best when combined with healthy eating and activity. But while roughly 15 per cent of people don’t respond to the current generation of GLP-1s, the success rate is drastically higher than through dieting alone, which for most people does not work for losing weight and keeping it off.
At the moment, however, the drugs’ reach is relatively small. In the U.S., an estimated nine million people have prescriptions to a GLP-1 drug, or a little over 2 per cent of the population. With a cost between US$900 and US$1,300 a month south of the border, the drugs are currently out of reach for most Americans without private health insurance.
The penetration rate is higher in Canada, where the cost of Ozempic without insurance is around $250 a month. Last year in Canada, more than 6.3 million prescriptions for Ozempic were filled out, up 80 per cent from the year before, making it the fourth-most-commonly prescribed drug in Canada, according to IQVIA Canada, a health care data and analytics company. The prescription count includes refills, so it is not a measure of the number of users of the drug. Two years earlier, Ozempic wasn’t even in the top 25 prescribed medications.
Even though the share of North Americans on GLP-1 drugs is small, the effects are already being felt. J.P. Morgan Research has estimated that GLP-1 users are already purchasing 8 per cent less food in the form of snacks, soft drinks and high-carb products to consume at home than the average consumer.
This has started to move the needle for food, retail and other companies. Last fall, Walmart Inc. noted a “slight pullback in the overall basket” of food and snack purchases, which it attributed to weight-loss drugs. Nestle SA and Danone SA have responded to the Ozempic revolution by rolling out meals packed with higher protein, since one effect of rapid weight loss is loss of muscle mass.
Other industries are also bracing for the upsides and downsides of a thinned-down population. Airlines are eyeing fuel cost savings – an analysis of United Airlines operations by Jefferies Financial Group Inc. found that if passengers each were 10 pounds lighter, the airline could save US$80-million per year on fuel costs. And gyms may benefit as GLP-1 users find they need to maintain muscle.
The biggest winners, however, are the millions of people with diabetes and obesity who are able to live healthier and longer with more confidence, said Mr. Cole. That in itself will be disruptive. “The effects on marriage are going to be significant,” he said.
GPT + GLP = GDP
It wasn’t all that long ago big thinkers in economics and technology fretted about prolonged technological stagnation. In the wake of the Great Recession, Tyler Cowen, a prominent American economist, lamented in his 2011 book, The Great Stagnation, that well-off countries had reached a “technological plateau.” Meanwhile, Peter Thiel, the Silicon Valley venture capitalist, summed up the sentiment when he observed in 2013, “We wanted flying cars, instead we got 140 characters,” referring to Twitter/X’s character limit at the time. The prevailing view was that technological progress had slipped into neutral. Innovation had stalled.
Mr. Cowen has changed his tune. Citing developments such as the record-setting speed with which life-saving COVID-19 vaccines were discovered and approved, advances in AI and Ozempic’s rise, he has declared the great stagnation over.
The question then becomes: How much will the current innovation boom power growth? To answer that, at least on the AI front, we need an answer to another critical question, said economist David Rosenberg, president of Rosenberg Research: “Over the next five to 10 years, what is the total addressable market as it pertains to AI? That is the $64,000 question.”
The lack of consensus explains the almost comically wide range of estimates for AI’s contribution to global GDP over the next decade. PricewaterhouseCoopers International Ltd., the professional services firm, predicts AI will boost global growth by more than 20 per cent. At the skeptical other end of the range, Daron Acemoglu, a professor of economics at Massachusetts Institute of Technology, argued in a paper published in May by the National Bureau of Economic Research that predicted productivity gains from AI are overstated and the technology will boost growth over the same time frame by only around 1.3 per cent
The impact of Ozempic on GDP is slightly simpler to estimate because the negative effects of obesity and poor health have long been known, measured and fretted over by health researchers.
Earlier this year, the economics team at Goldman Sachs quantified the drag on growth by drawing on academic studies that have shown people with obesity are less likely to have jobs and are less productive when they do. According to Joseph Briggs, Goldman Sachs’ senior global economist, obesity alone reduces per capita economic output by 3 per cent.
At the same time Mr. Christensen, the Danish economist, points to recent research in Europe that has found that workers with obesity take an average 3.8 more sick days per year than their thinner counterparts.
But it is in combatting wage discrimination that Mr. Christensen sees some of the most immediate potential economic benefits from Ozempic. Women with obesity earn roughly 10 per cent less than women of normal weight who are the same age, and have the same education and experience. If obesity in females could be cut in half, it could help close that gap, said Mr. Christensen.
“That alone would increase U.S. GDP by two percentage points, and that’s very substantial,” he said.
In its report, Goldman Sachs estimated that if 30 million Americans used GLP-1 drugs and 70 per cent of them experienced benefits, U.S. GDP could get a 0.4-per-cent boost. If the number of users doubled to 60 million and 90 per cent saw weight loss, GDP would get a 1-per-cent lift.
For his part, Mr. Rosenberg said the Ozempic effect may be real at the individual level, but he doubts it can meaningfully alter the economy’s trajectory.
“It’s a brilliant medical innovation that leads to better health and potentially longer longevity, but it will only change the lives of people who are diabetic or overweight. Generative AI will change everybody’s lives,” he said. “You wouldn’t be changing your quarterly GDP forecast because of Ozempic.”
He then sums up the Ozempic vs AI comparison more bluntly: “One is a small potato, one is a big potato.”
Yet each month brings hints these drugs offer even greater health benefits. Earlier this year, U.S. health regulators approved Wegovy to reduce the risk of heart attack and stroke in overweight patients with cardiovascular disease. Semaglutide has been shown to lower the risk of kidney and liver disease. In trials, GLP-1 drugs have slowed the decline of cognitive function in people with Alzheimer’s and lowered the rate of dementia in Type 2 diabetics. The drugs have also been associated with reduced alcohol consumption and curbed other addictions.
All these spinoffs are being put through tests and clinical trials. Even so, with a large pipeline of other GLP-1 drugs set to hit the market, human health on a massive scale could be at a turning point.
Fitter, happier, more productive. With an asterisk
Apart from how AI and GLP-1 drugs stack up against each other, both advances have hurdles to overcome if the hoped-for productivity boom can be achieved.
As impressive as ChatGPT is, it’s not yet clear that large language models will be enough to get AI “to the point where it will eventually reach and surpass human level intelligence,” said Mr. Toews. The newest version of the app, ChatGPT-5, is set to be released later this year or early in 2025, and it may feature autonomous agents capable of handling real world tasks such as ordering groceries online or booking travel. But those tasks may require a more fundamental breakthrough in AI technology.
Nor is it clear how governments will respond as AI advances. A bill making its way through California’s state legislature would require developers to run safety tests on large AI models and pave the way for California’s attorney-general to sue developers if a model causes a catastrophic event. And while the bill is backed by Canadian researchers Geoffrey Hinton and Yoshua Bengio, sometimes referred to as godfathers of AI, many companies and investors fear the legislation will stifle the sector.
Likewise, for weight-loss drugs to reach their full potential of transforming larger numbers of lives, GLP-1 medications will have to prove they are safe to be taken indefinitely.
For one thing, it appears some women are experiencing surprise pregnancies, suggesting the drugs may affect birth control or increase ovulation, yet GLP-1 drugs have not been tested on pregnant women.
Weight-loss drugs also have a long history of failures, including the appetite suppressant Fen-Phen, which was hugely popular in the 1980s and nineties until it was found to cause damage to heart valves and pulled from the market in 1997. While some people who use GLP-1 type drugs experience side effects such as nausea and stomach pain, nothing on par with Fen-Phen’s dangers has been detected.
A more pressing problem remains the steep cost of the drugs. By some estimates, if the U.S. made GLP-1 medications available to the 40 per cent of Americans with obesity, the annual tab for federal, state and private insurance programs would top US$1-trillion, or roughly one-fifth of all U.S. health care spending.
The price of Ozempic-like drugs is almost certain to come down though. In July, U.S. President Joe Biden went on the offensive against Novo Nordisk and Eli Lilly, demanding the companies “end their greed” and lower prices. The White House recently negotiated lower prices on 10 commonly prescribed drugs, and analysts suspect Ozempic may be included in the next round of price negotiations.
There is also a lot of competition in the pipeline. Several dozen rival weight-loss drugs are in development that could replace the need for injections, be more effective, have fewer side effects and cost less.
In the time since AI and Ozempic had their breakthrough moments, they’ve shared something else in common: being lightning rods for moral panic. AI will deepen racial bias; Ozempic is a diet cheat; AI will lead to mass unemployment; Ozempic is setting back body positivity; AI will amplify misinformation; Ozempic worsens health disparities.
The same dour mood tends to accompany most new innovations, no matter how positive they ultimately turn out to be. But if the twin growth enhancers of smarter machines and healthier humans do ignite a productivity boom, they may even provide the missing ingredients to get us closer to John Maynard Keynes’s dream of a 15-hour workweek by the year 2030, said Mr. Christensen.
“Are these breakthroughs going to help us live longer and healthier lives with more leisure time and more happiness? I think the answer is yes.”