Dropbox recently sparked concerns among its users with a new AI-powered search feature that shares data with OpenAI by default. The feature, part of the “Dropbox AI alpha,” is designed to enhance file searching using a ChatGPT-style conversational interface.
The discovery of this default-enabled setting led to user unease, despite Dropbox’s assurances on an AI privacy FAQ page.
Dropbox clarified that user data is only shared with OpenAI when the feature is actively used and assured that this data isn’t used to train AI models, being deleted within 30 days.
Dropbox CEO Drew Houston emphasized that no customer data is automatically sent to third-party AI services without explicit user action.
This feature is a part of Dropbox’s broader initiative to integrate AI into its services, like the previously announced Dash feature for universal searches across platforms. Dropbox further clarified that customers retain control over the use of these features, with data being sent to third-party AI only after explicit user interaction.
Despite these clarifications, the incident highlights the importance of clear communication regarding AI access to personal data and user consent in the evolving digital landscape.
Sources include: Ars Technica