This is Optimizer, a weekly newsletter sent from Verge senior reviewer Victoria Song that dissects and discusses the latest gizmos and potions that swear they’re going to change your life. This week’s issue is a special early edition tied to The Verge’s WWDC coverage. You can expect our next issue at its usual time next Friday. Opt in for Optimizer here.
Apple doesn’t like telling people what to do with their health. At least, that’s been true of its approach to the Health app and Apple Watch for the past 11 years. You can track various metrics, such as steps, workouts, and sleep quality. You can also view long-term trends — as in, whether your resting heart rate has risen or fallen in the past six months. Depending on the gadgets you have, you can also get alerted if you’re showing signs of sleep apnea, hypertension, and atrial fibrillation. But aside from nudging you to close your rings or perhaps consult a doctor if the Watch flags potential health conditions, it’s up to you to figure out what you should do with your data.
That’s a deliberate choice, as I wrote in a recent Optimizer about the Apple Watch. But while watching this year’s WWDC keynote, Apple’s approach stood in stark contrast to what I witnessed at Google I/O. At the latter, DeepMind CEO Demis Hassabis declared that the goal was to use AI to one day “solve all disease.” (A bold statement that requires context.)
Google believes so much in AI health that it’s willing to court the ire of Fitbit diehards by heavily promoting the Google Health Coach as one of the primary selling points for the new Fitbit Air. Other players in the health tech space — Oura and Whoop in particular, but also Garmin and Withings — have all added AI to their platforms, steadily rolling out new AI-powered features every few months. These new features can generate workout plans, log your meals, suggest macros, interpret your data, and, in some cases, even review your bloodwork if you choose to share it.
In contrast, health features weren’t center stage at this year’s WWDC keynote. Heck, watchOS 27 didn’t even get its own dedicated segment because the entire focus was on reintroducing the company’s vision for a smarter Siri and Apple Intelligence. There was a brief mention of perimenopause support in the Health app’s Cycle Tracking feature. After the keynote, Apple’s website noted a handful AI health and fitness updates in watchOS 27. Workout Buddy, an AI fitness feature introduced last year, will be available in Spanish, incorporate more fitness data in readouts, and work locally on the Watch when your iPhone isn’t nearby. There will also be more accurate algorithms for treadmill distance, plus other refinements in data syncing and GPS maps. Theoretically, you could possibly use the new Siri AI and Apple Intelligence features for health guidance in the same way you can with any other chatbot. For example, snapping a photo of your meal can now bring up nutritional facts. (The limits of Siri AI in this respect will be clearer as I test it.) But that too is something you’d have to figure out for yourself.
I wouldn’t be surprised if people interpret this as Apple falling behind on health AI and wearables. Before WWDC, Bloomberg’s Mark Gurman laid out a convincing case, reporting that Apple’s rumored Project Mulberry — supposedly an AI coach and doctor — was scaled back. He contextualized that alongside several longtime Apple Watch and health executives recently leaving or retiring. Meanwhile, screenless trackers like Whoop, the Fitbit Air, and Oura Ring seem to be gaining in popularity. Apple’s last major lineup shakeup was in 2022, and since then it’s been relatively iterative updates. As far as many gadget nerds and industry analysts are concerned, Apple needs to wow everyone or accept that the Series 12 will fall by the wayside.
I get this line of thinking, but I’m not quite as convinced based on my experience testing many AI health coaches. I don’t necessarily see this lack of AI health news as a bad thing. I’ve said it before, and I’ll say it again: The vast majority of AI health features are doo–doo dogshit, tedious to train, and require so much of your personal data to be at all effective that it feels ominous.
Of the health coaches I’ve tested, Google Health is the most well-rounded, but the bar is deep in the ninth ring of Dante’s inferno. That’s not great when the stakes with health are much higher than an AI assistant mucking up a DoorDash order. The most responsible use case for any health coach thus far is as a “helper tool” between doctor’s appointments to help patients stay on track and have more productive conversations with their healthcare providers. Rushing to implement health AI might suit investors, but from talking with fellow reviewers and users, most find one or two things useful. The rest of those conversations are a laundry list of complaints and wish lists for improvements. I appreciate that this is an ongoing process, but so far, I feel a Pavlovian urge to groan every time I hear about yet another AI health feature.
These might come off as fighting words. They’re not. After interviewing several doctors and researchers for Optimizer, it’s clear to me that AI has been a major contributor to health breakthroughs for decades. It has a role to play in the future of healthcare. Vaccine discoveries, for example, have benefited in recent years from AI streamlining several steps of the development process. What we haven’t nailed is how generative AI improves the consumer health and fitness tracking experience compared to a Google search.
Could the Health app use a redesign? Sure. It is a bit cluttered, but show me a health and fitness tracking app that isn’t. (I’ll wait.) Do the people yearn for screenless trackers because of endless data fatigue? Big yup, but current AI health insights are rarely more than glorified, regurgitated summaries of obvious charts anyway. I can’t lie. If Apple doesn’t have a unique, genuinely helpful take on this space, I’m relieved that we heard nothing. I’d rather not test an underbaked Siri-flavored health coach this year.
Would it have been more exciting if Apple had surprised everyone with an innovative, never-before-seen, completely private, and utterly game-changing implementation where AI flips the script on healthcare as we currently know it? Duh. But that’s something I’m willing to wait for.
I’m currently experimenting with anything and everything under the sun to help with some health issues I’ve been dealing with. Despite my innate skepticism, I’m always willing to keep an open mind in case I’m proven wrong. Yet, it becomes soul-crushing when nearly every attempt aimed at consumers feels like a thinly veiled, stapled-on cash grab. And this is coming from someone with annoying but not life-threatening conditions. I can imagine how much more frustrating the current limitations of consumer AI health tech can be for people suffering from serious chronic conditions or terminal illness. Or disorders that can be exacerbated by AI sycophancy or hallucinations, like eating disorders or hypochondria. A good AI health implementation has to be useful for those of us who are not in the best of health, too.
I can understand the impulse for companies to “not miss the boat.” But taking a moment to zoom out at the current state of consumer health tech? It’s a hot mess. Most users aren’t happy, and now more than ever, it feels like AI is ironically exacerbating health data fatigue rather than solving it. Perhaps we could all do with Big Tech taking an extra beat or two to think things through.
