Starlink's Revised Privacy Policy Opens Door to AI Data Harvesting

Starlink users worldwide are now a potential data source for artificial intelligence projects. A January 15 update to the satellite internet service's privacy policy explicitly grants the company permission to use customer information to train machine learning and AI models. The change, which applies automatically to all users, also allows for sharing this data with third parties for their own AI training purposes.

Upon logging into their accounts, customers are now met with a notification stating, 'You allow your data to be used to train AI models.' While Starlink asserts that individual browsing history and geolocation data will not be shared with AI models, the policy broadly covers contact details, performance metrics, billing information, and even 'communication information, such as audio, electronic or visual information.'

William Budington, a technologist with the Electronic Frontier Foundation, criticized the move as part of a broader industry trend. 'It's part of this whole rush to just throw everything into the data-driven machine-learning nexus,' he told CNET, expressing concern for user privacy.

The policy revision precedes by weeks the announcement that SpaceX, Starlink's parent company, is acquiring xAI, another venture under Elon Musk's portfolio. With over 9 million global subscribers, the potential data pool is vast.

Users who wish to opt out can do so through the Starlink app or website. After logging in, navigate to Privacy Preferences in the account portal and deselect the option permitting AI training. Security experts, however, recommend additional measures. Using a trusted VPN encrypts communications, preventing Starlink from accessing the content of online activity, though it must be configured on each device or router for full protection.

Source: CNET