It’s difficult to get a bead on exactly how the AIRO wristband works, and this makes cross-checking its scientific validity all the more difficult. I spoke with the company’s original CEO and co-founder, Abhilash Jayakumar, but he wouldn’t shed any light on the technology beyond pointing me to public statements he made when he was a member of the Airo team. But Jayakumar did explain why he left Airo Health: “The high-level reason is that I had a fundamentally different vision for the company than my co-founders, who were united on theirs,” he said.
I also interviewed Airo Health’s current CEO, Naman Kumar, another one of the company’s original co-founders. Try as I might, I couldn’t get Kumar to explain in detail how the AIRO wristband tracks calories with a spectroscopic sensor, but he did say the sensor is more of a “nutrition sensor” than a “calorie sensor.”
Based on Kumar’s brief overview and already published information about AIRO, it seems the wristband uses a miniaturized spectroscopic sensor similar to the technology once under development by C8 MediSensors, a now-defunct company that was working on non-invasive blood glucose monitoring for diabetics (the company’s Director of Optics and System Engineering, Ueyn Block, now works for Apple).
In a nutshell, Airo’s system is similar to a teeny-tiny flashlight and camera: An LED array shines different wavelengths of light through the skin, while a highly sensitive photo detector determines which wavelengths have been absorbed, and which have been reflected. The system detects the optical footprints of what Airo Health describes as metabolites, and from these, the AIRO algorithm can estimate calorie intake.
Kumar wouldn’t call AIRO a system that primarily senses blood glucose levels. He would only tell me that “glucose-sensing is one way to derive calories. It’s one of the ideas that has emerged from theoretical speculations, but it doesn’t need to be that.” He did say, however, that the AIRO system has greater success in tracking carbohydrates than other types of food: “Mostly stuff that is rich in carbohydrates—rice, bread, that kind of stuff. But proteins and fats need more time and optimization,” he said.
As for what compelled Airo Health to suspend its crowd-funded pre-order campaign, Kumar says his team hit two road blocks: It couldn’t produce reliable, scalable results beyond the lab environment, and it still has work ahead in making sure the wristband sensors always make reliable contact with the skin.
“We hadn’t done enough testing to be able to say that, ‘Yeah, not only we in our lab can use it, but you as an outside party, a third-party consumer can use it too,’” Kumar said. “We need to make sure that it works on different kinds of people. You and I are different. There are differences in how your body processes food, in how your body changes every time you eat something.”
Kumar told me that AIRO’s “expected entry to market” is mid-2015. But Ries Robinson—whose current company, Medici Technologies, develops predictive algorithms for the fit-tech industry and other sectors—said that AIRO, as it’s been described by company founders, is beyond the scope of feasibility, just like GoBe.
Robinson’s takeaway is that AIRO is focused on blood glucose, just like the GoBe system, and this is already an invalid calorie-tracking model, as it doesn’t account for dietary proteins and fats. But even more damaging, if we assume AIRO is “just” a blood-glucose sensor, it still wouldn’t have enough science behind it to produce reliable results, he says.
“Is it plausible by using one or two light emitting diodes—which looks to be the case with AIRO—could you measure glucose in the tissue with that type of optical platform? The answer is no,” says Robinson.
“We probably raised and spent $70 million trying to measure glucose non-invasively in the tissue,” Robinson said of his own experience in the field. “We had two Ph.Ds in tissue optics, and a group of individuals who were extremely good at quantitating spectroscopy, and I think we came close to realizing a device, but the hardware to do that is very sophisticated. It still has a price point even today of $5,000. There’s no way you’re doing it with two LEDs and a wristwatch.”
The case for doing nothing
Consumer tolerance for inaccuracy is already pretty high in the activity-tracking wearable space. The estimated 50 percent of users who actually stick with their wristbands seem comfortable with the fact that basic accelerometer-based reports are inconsistent from platform to platform—that if you wear three different activity-trackers on your arm, each will report different step counts and calorie-burn numbers.
What’s more important to consumers, it seems, is that calorie burn reports prove accurate within the context of their own closed platforms. And to this point, the wristband manufacturers typically publish some type of claim regarding the accuracy of their step counts and calorie-burn algorithms. Here’s one from Fitbit. And here’s one from Basis Science.
But calorie-intake tracking, based on everything we’re learning from experts, can’t even approach a minimum level of accountability. So is it even worth exploring, especially with all the negative scrutiny that’s been focused on logbook methods like MyFitnessPal and quasi-scientific methods like those mentioned in the article? Basis Science, for one, has decided to avoid calorie-intake tracking entirely—at least until the proverbial holy grail of fit-tech emerges for real.
“Today, the only ways to do track calorie intake, are, unfortunately, manual,” Basis CEO Jef Holove told me during a product demo in October 2013. “We know the user engagement just isn’t there. Consumers will say they want to do it, but yet our real-life habits just aren’t there. When we feel we’ve discovered some way of doing calorie intake that makes us feel better about it, then perhaps that gets elevated.”
In other words, Basis would be among the many companies interested in the wearable fit-tech version of cold fusion. But it also knows that today—with available technology, in a universe that must conform to the laws of physics and can’t indulge in science-fiction dreams—automatic, accurate, sensor-based calorie-intake tracking is beyond the scope of feasibility.
This story, "Wearable snake oil: The search for automatic calorie-intake tracking in fit-tech wristbands" was originally published by TechHive.