AMAZON’S NEW HEALTH tracker, the Halo, is a small, display-free device fitted into a fabric band that you wear on your wrist. Unlike many fitness trackers that simply count your steps, the Halo collects additional health data that is shockingly intimate. Most notably, the app analyzes pictures of you in your undies to measure your body fat percentage. Also, the Halo’s microphone records snippets of your conversations to interpret the emotions conveyed in the tone of your speaking voice.
I’ve worn the Halo for a month, and it lacks a lot of features I’d want in a fitness tracker, like GPS for tracking your runs. In fact, I’ve also found that most of the data the Halo does analyze don’t seem worth the hassle of collecting it. I’m not sure it’s a great idea for any company, let alone Amazon, to normalize the process of full-body strip-down scans for its customers.
I’m not the only one who thinks so; the requests for underwear photos and the on-device mic have raised many privacy concerns. Despite these complaints, the Halo’s voice-tone tracking feature intrigued me. I wore the wristband through a stressful pandemic holiday with two small children. After two weeks with my nuclear family in the claustrophobic confines of our home without childcare help, parties, or any rest, the Halo began registering voice tones it flagged as increasingly irritable. It made me wonder … if I actually am getting irritable, what the heck am I supposed to do with this data? I decided to find out.
The Halo assesses your emotional state with a small microphone on the side that you can turn on or off. You teach it to recognize your voice by reading prepared phrases, and throughout the day it records small snippets of conversation and rates them for their “positivity” and “energy.”
These emotions manifest in the app as four different emoji-like faces: A laughing yellow, a smiling green, a weary purple, and an angry red. The app saves notable moments throughout the day, telling you how you felt at different times: irritable at breakfast, amused at dinner, or content in the hour before bedtime.
You can also force it to monitor you during important conversations by holding the button on the side of the Halo. Or, open the mobile app and press the Live tab to watch it monitor your tone of voice in real-time. In my household, at least, there is no better way of dispelling marital conflict than by saying, “Let’s have a heavy conversation so I can monitor it with this app.” I tried it a few times, but my husband and I ended up just crouching over my phone, talking in funny voices to see how the app would register our vocalizations.
In fact, the Halo was rarely able to recognize when I was mad at all. One possibility that Ty Tashiro, who holds a Ph.D. in psychology and is the author of The Science of Happily Ever After, brought up over the phone was that I may express my feelings in gesticulations or other nonverbal communication, and not in the tone of my voice.
If you start from the (perhaps generous) assumption that the Halo’s emotional data is accurate, it’s missing a lot of contexts—specifically, those nonverbal cues and behavior traits unique to each individual. “Only you know the receiver and the sender and how they’re interpreting all those variables—content, context, nonverbals, as well as tone—to say, ‘Oh, my wife is being sweet in a sarcastic way’, or ‘She’s irritated and using a sweet voice,’” Tashiro says.
The Halo’s emotional rating system of colored smiley faces is not without precedent. They bear a resemblance to the work of renowned therapist John Gottman, who observed couples and developed a coding system to rate their positive or negative behaviors with each other.
But Gottman’s research goes a lot deeper than the Halo. It examines real-time interactions between the couple and includes facial expressions and gestures in addition to the tone of voice. In fact, of the five positive behaviors and 10 negative behaviors that Gottman measures, some of the most telling negative ones are silence and withdrawal (what Gottman calls “stonewalling”). If two Halos could work in tandem, measuring real-time discrepancies between the volume, tone, and pace of two people to indicate bad interactions, there might be a lot of interesting possibilities. But they cannot.
Tashiro also made an important point: If the Halo is providing too much negative feedback, it might cause you to avoid bringing up important issues in conversation. “Just because you get dinged that your tone is irritable or angry doesn’t mean you should stop,” he says. “Maybe your partner is being a jerk. As [social psychologist] Dr. Ellen Berscheid would say, sometimes communication isn’t the problem. The problem is that the partner doesn’t like what they hear.”
To help interpret the data it collects on the tone of my voice, Amazon’s app partners with outside organizations to create labs, which are small exercises meant to increase mindfulness in the wearer. In one lab, I had to ask myself the nonsensical phrase “What is listening?” before I talked to anyone each day, to become a more “conscious listener.” I found these sorts of exercises unhelpful when it came to interpreting my own data, so I turned to marriage therapist Terry Real.
I was expecting Real to be extremely skeptical about newfangled devices that use cloud-based AI to interpret biodata. After all, human relationships are much too complicated to benefit from these, amirite? But he surprised me right off by disclosing that he uses similar devices in his practice. “HeartMath has a gizmo called the Emwave,” Real said. “You clip it to your ear or put your thumb on it and turn the light from red to blue to green by doing centering practices. When it’s green, you’re no longer in a reactive state. I have used this for decades with highly reactive partners.”
As it turns out, real-time biofeedback is useful—to an extent. These “tools are a way of tracking when people lose their shit,” Real said. “It helps them get recentered in a more adult way. But beware of what I call objectivity battles, where the data proves you’re this or that.”
You need an experienced therapist to translate the data into action. Because Real is a therapist, I found myself sharing my data with him and, for context, telling him about a source of conflict in my marriage. I grew up outside a big city on the East Coast, while my husband is from a small town in the south. Because of the different social conditioning we experienced growing up in these disparate regions, he occasionally finds the way I talk—the speed, the tone, the sheer volume of interesting facts or things to remember or requests that I spout off in short periods of time—to be aggressive, overwhelming, or confrontational.
I’ll give him some of those points, but the Halo also provides evidence that, in general, I am mostly cheerful, high-energy, and rarely irritable or angry. There’s no problem, right? Real disagrees.
“I like to say the answer to the question who’s right and who’s wrong is ‘Who cares?’” Real said. “The feedback is good, as long as you don’t get hypnotized by the power of objectivity. I don’t want technology to be more important than your humanity. If your husband thinks you’re being aggressive, your husband wins over the Halo.”
“Objectively, you’re not being irritable. He’s making that up,” Real continued. “Subjectively, they talk slow and gentle where he’s from, and you’re not, so could you please talk more like what he’s used to? Why not? You’ll try. It only goes so far, and he’s going to have to accept how far you get. That’s how real couples operate.”
After my month of testing, I couldn’t wait to take the Halo off. With its paltry selection of features, I just didn’t find it useful as a fitness tracker. It also doesn’t motivate me to move as much as it should. The Halo encourages the wearer to collect 150 activity points per week, but I was accumulating that many in one day. And the display-free design didn’t help me make it to any meetings on time; for that, I really need an actual watch face.
As for those 3D, full-body skivvy scans, they aren’t any more reliable or consistent than just getting on a smart scale. Just like with the scales I’ve tested, the Halo’s body scans swung widely from 26 percent body fat to 33 percent, depending on everything from what I was wearing, the time I did the scan, or how many mini pizza bagels I’d had for dinner.
Still, as invasive as it was for a wristband to listen to me talk and interpret the tone of my voice each day, the $100 Halo changed me. I had to do the legwork myself, but it did spark change and improvement in one of the most important relationships of my life. Fitbits and Apple Watches may help me stay in shape, and a simple watch would help me start work on time, but the Halo deserves some small commendation for strengthening my marriage in the middle of an oppressively lonely global pandemic.
If you buy something using links in our stories, we may earn a commission.
This helps support our journalism. Learn more.
Special offer for Gear readers: Get a 1-Year Subscription to WIRED for $5 ($25 off). This includes unlimited access to WIRED.com and our print magazine (if you’d like). Subscriptions help fund the work we do every day.
- 📩 Want the latest on tech, science, and more? Sign up for our newsletters!
- Yes, Cyberpunk 2077 is buggy. But mostly, it has no heart
- A racecar crash from hell—and how the driver walked away
- Apple’s app “privacy labels” are a big step forward
- These 7 pots and pans are all you need in the kitchen
- The race for a Covid vaccine was more about luck than tech
- 🎮 WIRED Games: Get the latest tips, reviews, and more
- ✨ Optimize your home life with our Gear team’s best picks, from robot vacuums to affordable mattresses to smart speakers