Strava will soon use AI to weed out cheaters from its leaderboards
Strava has announced it will soon use machine learning to automatically flag irregular or improbable activities uploaded to the platform, as part of a slew of new features arriving before the end of the year.
The platform’s leaderboards have been cleaned up considerably over the last few years, with most users tracking with a running watch or smartwatch and automatically syncing, yet the likes of Strava Segments could still be fudged with an edit of a workout inside the app.
It’s not clear how successful this will be in practice – Strava simply states that its system has been trained on millions of activities – but the hope here is that the leaderboards will be more fair and accurate.
Given that the community flagging system currently in place often sees activities from high-profile users spammed and pulled – like with cyclist Tadej Pogačar, whose upload from stage 15 of the Giro d’Italia was recently targeted – we’d say it’s a long overdue rethink.
And it’s not the only long-requested feature coming soon.
Dark Mode & Night Heatmaps
Strava also announced that ‘Dark Mode’ will also be arriving later this summer on iOS/Android versions of the platform, giving users the option to match the app’s background to their phone’s settings.
Perhaps the more interesting new addition, though, is the all-new ‘Night Heatmaps’ (shown below).
Announced as part of Strava’s new ‘Build for Her, Build for Many’ initiative, the heatmaps filter activities that have taken place between sundown and sunrise to give users a better idea of the areas that are well-trafficked after hours.
It’s a superb and sensitive new feature – and one that only really a platform with this many users (125 million, Strava says) can pull off to any meaningful effect.
Away from using it as a safety feature, it should also be handy as a way to mix up your runs or rides during those darker nights; we’re much less adventurous in the winter, but Night Heatmaps has the potential to unlock some new routes.
Family Plan & Athlete Intelligence
Elsewhere, the company also announced a new ‘Family Plan’ subscription and a new ‘Athlete Intelligence’ feature.
With the new family plan, users can join up with up to three other people (interestingly, you don’t have to share an address, so this is very much open to any group) and receive the paid-for benefits.
It’ll begin rolling out to some countries later this summer and globally before the end of the year, with specific pricing details TBD. Strava did say in its release, however:
“Strava’s newest annual subscription option offers the best value for groups (up to four), with a discount off the regular subscription price for each member.”
Those other AI-powered features – tagged as Athlete Intelligence in the app – will also be coming soon, and tap into training data to deliver more meaningful insights.
At first glance, it appears akin to what we’ve seen from other tailored AI betas in fitness apps – particularly Whoop Coach. However, given that Strava can pull data from tons more wearables and sources, we’re hoping for some next-level, proactive insights here that can add some deeper context to our workouts.
We’ll be playing around with these updates once they begin rolling out – stay tuned for more.
