AI’s influence on Activity Trackers

AI’s influence on Activity Trackers

by Christian Bolorinos

Posted on June 14, 2020

An industry that I have been intrigued by for some time is Activity Trackers. Activity Trackers can loosely be referred to as “wearable computers” or “smart watches” that monitor your body to help you achieve your health goals. Despite the name they can take a variety of forms. The technology started with heart trackers in the 80’s and bicycle computers in the 90’s. It has now expanded to include mostly watches but also shoes, rings and earbuds. The latter currently accounts for 50% of the market; more so than any other form of wearable and up from 27% only two years ago. This is largely due to the fact that ears transmit clearer data than watches. The future will be seeing a divergence of wearable activity trackers, from coffee mugs to baby onesies, all of which have their own benefits and drawbacks. The future is also likely to see an increase in the data these sensors are able to collect. Right now it is little more than fitness habits, sleep, calorie intake and heart rate. However sensors for new functions are likely on the horizon, including the chemical makeup of one’s urine.

The industry is currently dominated by Xiaomi, Apple and Fitbit (in that order) holding the largest share of current shipments, and Nike and Garmin trailing behind them. Fitbit could probably be considered the pioneer of the bunch, and the only one who didn’t move laterally. After its largely unsuccessful attempt at entering into wearable technology with Google Glass, Google is trying to enter the market by purchasing Fitbit. Google is a pioneer in using user data to tailor their experience, so this acquisition is likely to make Fitbit more personalised to its users. When Google announced this plan in 2019, it led to an increase in both Google and Fitbit stocks. However, there has been some pushback, as users are less likely to trust Google with their data.  

The public’s general mistrust over handing their information to major corporations is one of the main criticisms of the industry. The data collected could be sold to health insurance providers for instance, who may use it to justify changing their rates. Or it could be used to target ads to users much more directly. As data is one of the most highly valued commodities in tech today, this is likely to happen in some form at some point, so the public’s concern is largely justified.

Currently, the industry currently leaves a lot to be desired, and has until now yielded poor results. In fact, one study found that they currently contribute little to weight loss, and actually made wearers lose less weight than their non watch wearing counterparts. This may be because people who are told how many calories they burned are more likely to treat themselves to an extra slice of cheesecake after their workout. When it comes to tracking sleep habits the results are somewhat more promising, but not by much; resulting in only slightly better approximations of sleep habits by the wearers.

There is some speculation that these problems are caused by the incredible diversity of human bodies, and the lack of a standard form of measurement. However, while the technology is currently a mildly informative gadget at best (and a serious breach of privacy at worst), there is some indication that AI could change the industry, and revolutionise personal health. AI could detect habits, and personal traits that would otherwise distort the data. It could also find trends in user segments and use them to update the way information is collected. For instance, if you are walking up a hill you burn more calories than if you are walking on a level surface. Current activity trackers don’t factor this in, but with AI they would be able to do so. There are already some companies using AI in activity trackers to suggest more tailored workouts. Some activity trackers are moving the industry into health care by keeping track of users' heart rates and detecting anomalies that could lead to heart attacks.

Some recent startups have moved into psychology applications as well. The startup Limbic is collecting heart rate data through activity trackers to determine what makes them angry, nervous or depressed. AI is then used to gain insights into what triggers these changes in heart rate, which could help their users address or avoid these triggers. Once again, the specter or privacy violations and data breaches looms. But in this case one could also argue that the potential gains are enormous. Psychology up until now has suffered from a lack of quantifiable data, which contributes to a climate of unreplicable experiments with unverifiable conclusions. Limbic is setting out to change this, and has a very real chance of succeeding.

With innovations like these, the privacy concerns may come into a different light. I would be uncomfortable with Google knowing how many days I spent eating too much, suffering from insomnia and procrastinating going to the gym. However, if I’m having heart problems I would like my doctor to know as soon as possible, even if that came at the expense of Google obtaining the same information and selling it to my insurance provider. Similarly, I would consider the risks of a data breach if I could gain real, actionable and data based insights on what kind of activities trigger anger, depression or anxiety in me.

This has the potential to be misused for advertising or manipulatory purposes. However, it also has the potential to increase human life expectancy and general wellbeing if used constructively. The sheer quantity of data collected from Google has already given us a plethora of insights into the human mind. Combined with artificial intelligence, activity trackers might give us the opportunity to obtain similar insights into psychology and medicine. If this data was available to the right people it could result in newer and more effective forms of therapy or hospitalization. These devices could also empower users to take their health into their own hands, giving them an understanding of their own bodies and minds that is not currently accessible to the public.

As is generally the case with AI, this new application of technology requires careful consideration. Companies should be sure that the data is being used in a constructive and empowering way, and government regulations should be put in place to assure that this data is never used to manipulate or exploit users. Only time will tell if this is a problem, a solution, or as is usually the case in technology, a complex blend of both.

Get in touch