Strava blocks sharing of fitness data with other apps (19459000)
Strava informed its users and partners recently that new terms for their API restrict data that third-party applications can show, refraining from replicating Strava’s look and placing a ban on data being used “for any model learning related to artificial intelligence or machine learning, or similar applications.”
The policy is in effect as of November 11, even though Strava has posted about the change on its own blog, which is dated November 15.
The updated API agreement, which goes into effect on November 11, introduces three major changes that will give Strava users greater control, security and a consistent user experience.
Third-party apps are now required to adhere to stricter privacy standards. They can only show Strava activity data for a specific user. Users will still have access to Strava data in apps connected to the platform, but there may be some differences in the way this data is displayed.
Data Use Limitations: Our terms prohibit third parties from using data obtained via Strava API in artificial intelligence models and similar applications.
Protecting the Strava Experience – Additional terms have been added in order to protect Strava’s unique look, feel and functionality. This will help users distinguish between Strava (and third-party platforms) easily.
In a message sent to The VergeStrava’s VP of Communications and Social Impact Brian Bell stated, “We expect that these changes will only affect a small fraction of applications (less than.1%) on the Strava Platform. The overwhelming majority of use cases that are currently allowed are still permitted, including coaching platforms that provide feedback to users and tools which help users understand their data and performances.”
Below you can read the full update from Strava.
Ray Maker aka DCRainmaker points out that Strava is widely used as a middleman for data collected by wearables and transferred to other apps or services. Maker says that “every partner I’ve spoken to in the last 24 hour has been completely blown away by this change”, with their only notification being an email sent on Friday, giving them a 30-day warning to make any changes.
Maker, who is a Strava user, says that “Athlete intelligence” summaries generated by Strava “tell me vaguely positive, but usually wrong.”
TrainerRoad, an app that uses Strava data to guide users to “Effective, Scientific-Backed Exercises,” has already commented on this change. Nate Pearson wrote on the forum that “this was crazy news for us, too.” He also noted that Garmin Connect users could use it instead, and that the team already worked on integrating other platforms such as Zwift or Wahoo.
Brian Bell, Strava VP Communications and Social Impact
We wanted to give some additional context about the changes to our API Agreement, and the impact on our users and developers. We anticipate that these changes will affect less than.1% applications. We notified those affected last week.
Enhancement of privacy and user control
Privacy and user control is at the forefront of what we do. We are committed to changing our API practices in response to changes in regulatory requirements and user expectations. This includes limiting the ability of third-party apps to display a user’s information in ways that the user may not have expected. We want to address situations where users connect with a third-party application and are unaware of the fact that their data will be surfaced to other users, not only for their own use or visibility, but to all users (for instance, in a heatmap or public feed). The latest API updates address this scenario, and provide a more consistent framework to Strava user data.
Training AI Models
We believe that AI has the potential to transform the athlete’s experience. Whether it is delivering more personalized insights, helping you reach your goals or generating route or workout recommendations, there are countless possibilities. Innovation in this area must be handled with care and a focus on user control. We are committed to implementing thoughtful AI features that prioritize user control.
Third-party developers might not be as deliberate in their approach to training AI model and we believe that the best decision is to prohibit this use of data from Strava users. We had already prohibited the use of Strava data for model training and development in our previous terms, but we have made this more explicit due to the increased activity in this area.
No Impact on Most Developers
We acknowledge that our platform thrives due to the creativity and dedication shown by third-party developers, who create tools to complement Strava and extend its capabilities. We remain steadfast in fostering this ecosystem. We expect that these changes will only affect a small fraction of applications (less than.1% of all Strava platforms) – the vast majority of existing uses cases are still permitted, including coaching platforms that provide feedback to users, and tools that help them understand their data and performances.
Strava’s forums are not a place for dissent. The company states that “…posts asking or trying to have Strava reverse business decisions will not permitted.”
Update for November 19th Strava has added a statement.
