Pinterest has updated its privacy policy to reflect its use of platform user data and images to train AI tools.

A new clause, published this week on the company's website, outlines that Pinterest will use its patrons' "information to train, develop and improve our technology such as our machine learning models, regardless of when Pins were posted." In other words, it seems that any piece of content, published at any point in the social media site's long history — it's been around since 2010 — is subject to being fed into an AI model.

In the update, Pinterest claims its goal in training AI is to "improve the products and services of our family of companies and offer new features." Pinterest has promoted tools like a feature that lets users search by body type and its AI-powered ad suite, which according to Pinterest's most recent earnings report has boosted ad spending on the platform. The company is also building a text-to-image "foundational" AI model, dubbed Pinterest Canvas, which it says is designed for "enhancing existing images and products on the platform."

The platform has stressed that there is an opt-out button for the AI training, and says it doesn't train its models on data from minor users.

...

Soon after we reached out to Pinterest with questions about the update, we were contacted by a spokesperson who insisted that the update wasn't newsworthy because the update simply codifies things Pinterest was already doing. Later, the company provided us with an emailed statement.

"Nothing has changed about our use of user data to train Pinterest Canvas, our GenAI model," read the statement. "Users can easily opt out of this use of their data by adjusting their profile settings."

Pinterest was already training its AI tools with user data, as the company touches on in this Medium post about Canvas, but the practice is now codified in the platform's terms of service.