I ran a half marathon with an AI trainer and was not impressed
Blog

I ran a half marathon with an AI trainer and was not impressed


I did something very dumb last weekend: I decided to challenge the limits of artificial intelligence and my own body on a 13.1-mile course to glory. 

My body held up fine — but the AI couldn’t quite keep the pace, serving as yet another reminder that current AI tech still hasn’t caught up to the hype. 

I had been testing out Vi, a set of $249 Bluetooth running headphones with its own built-in AI assistant and biometric tracking features. The headphones are comfortable on long runs with a neck harness design and in-ear buds for great sound — but the AI assistant is the main attraction here. It’s like Siri and Alexa with a fitness-focused twist: Vi is designed to train its users to be mile-churning super runners. 

I wanted to really push Vi to see what the AI was capable of outside of basic runs, and I was craving a new fitness challenge to spice up my workouts — something that could take me from a casual runner logging intermediate distances to a true blue marathoner, or at least help me manage my pacing. 

After a convoluted series of events in which I was offered a potentially illegal entry to the Brooklyn Half Marathon a week before the race, I found my adventure: I decided to run my own 13.1 miles in the Prospect Park Loop with nothing but the AI headphones to guide me, using Vi for a crash training course to prep in less than a week.

Vi checks all the boxes seemingly required of an AI assistant: it has a catchy name, a disembodied-yet-reassuring voice, and, unfortunately, not nearly as much functionality as you’d expect from your fancy new gadget-based friend. 

Vi doesn’t offer much more than other running apps I’ve used: It tracks the distance you run, measures your heart rate, and offers some realtime coaching direction to fine-tune your step rate to find your ideal pace, which it calls your “Comfort Zone,” — but it leaves much to be desired as a next-gen personal trainer. It currently has no dedicated feature to set specific goals, so users prepping for races like me have no guide to train for big events or set more defined goals than just fine-tuning their running style. 

Vi and me, just getting to know each other.

Vi and me, just getting to know each other.

Image: lili sams/mashable

Putting Vi through her paces

I decided to take the challenge on a Monday with the race on Saturday. That gave me four days to get ready for the half marathon. 

A quick note for you, dear readers: Don’t try this at home.

Runner’s World suggests committing between 10 to 14 weeks to train for a half marathon, especially for novice racers. I’m active, and run roughly the equivalent of 10 miles a week between jogging and Muay Thai training, but I had never run a road race or jogged over seven miles in one go before this adventure. This was a capital-B Bad Idea, to which I committed for the noble cause of #content and living my best #runlife.

Vi didn’t give me much to go on during the pre-race training sessions. The AI noted we had taken our longest run together, but since it was already accustomed to my step rate, it didn’t give me guidance other than when I left my “comfort zone” and at the end of the workout when I hit my target distance.

When I took to the park loop for the big run, then, I wasn’t sure what I could expect from Vi or myself. I synced up the Vi app for a 13.1-mile Distance Run, put in the earbuds, and took off.

As I ran, I noticed again that Vi wasn’t providing me with insights on my running technique, which is what you’d pay for with a personal trainer or coach. After each mile, Vi told me my split time and distance covered, which was helpful — but I was getting the same data from Map My Run, a GPS tracking app that doesn’t boast AI functionality.

I was able to track my heart rate and speed — but those measures couldn’t give me much to go on in the moment. 

I made it!

I made it!

Image: screenshot/vi

I finished the race, but I felt like Vi left me hanging. My last three miles were actually faster than the 10 before them because I wasn’t sure how to pace myself, which is one of the first orders of business for a trainer to establish in the lead-up to a long race. 

AI assistants aren’t quite ready — yet

The issues I had running with Vi reminded me of the problems I’ve had using other voice-based AI assistants. The novelty of being able to talk with an otherwise inanimate object loses its luster quickly when that conversation goes nowhere, as anyone who has struggled to pry more than just a basic Google search out of their phone’s AI can tell you. That frustration goes double when you’re out of breath after running for an hour.   

Vi didn’t launch as a fully-realized personal trainer, which reminds me of another notable AI assistant: Samsung’s Bixby. The much-heralded AI got its own dedicated button on the Galaxy S8, but when the phone launched in April, Bixby wasn’t ready for the party. 

Bixby’s failure to launch was a misstep for Samsung, and just the most recent evidence that the current reality of AI isn’t exactly on par with the hype surrounding it. 

That doesn’t mean Samsung’s assistant will ultimately fail, especially considering the world of AI it will help to shape. Many of the most exciting recent developments from Google and Microsoft focus on practical AI applications, which aim to bring the sophisticated systems to just about every one of their products and programs. 

Even Vi looks like it will much more to offer in the near future. The company’s reps tell me the AI just got some essential upgrades, including new insights and prompts to give users during runs to take advantage of its capabilities. Most importantly, a dedicated race prep functionality is due for release later this summer. Those tweaks could make Vi a fully actualized trainer, rather than just a cool pair of headphones with some fit tracking features.  

Once that race training feature has a chance to find its legs later this year, maybe I’ll give Vi another shot. I only finished half a marathon with her in my ear, after all — we’ll have a much better chance to get comfortable with each other over a full course and 26.2 AI-aided miles.

Https%3a%2f%2fblueprint api production.s3.amazonaws.com%2fuploads%2fvideo uploaders%2fdistribution thumb%2fimage%2f1944%2f6e0bed34 7886 4c07 85dc 0baf265cd33b


May 31, 2017
Comments

Leave a Reply

Your email address will not be published. Required fields are marked *


9 + 2 =