The early architecture of the modern AI era was built on a foundation of borrowed data, often harvested from the digital commons without the explicit consent of those who inhabited it. In a belated reckoning for this practice, the facial-recognition firm Clarifai has confirmed the deletion of approximately three million photos sourced from the dating site OkCupid, along with the algorithmic models trained upon them.
The data transfer occurred in 2014, a time when the ethical guardrails for machine learning were largely theoretical. OkCupid provided the images to Clarifai to help refine its facial-recognition capabilities, a move that the Federal Trade Commission (FTC) later determined breached the dating platform’s own privacy promises. While Clarifai itself was not accused of legal wrongdoing, the move to purge the dataset follows a late-March settlement between the FTC, OkCupid, and its parent company, Match Group.
The settlement carried no financial penalties, but its impact is felt in the forced destruction of the intellectual property derived from the photos. For Clarifai, a Delaware-based company that has become a significant player in the computer vision space, the deletion marks the end of a controversial chapter. It serves as a reminder that the "move fast and break things" ethos of the previous decade is increasingly colliding with a regulatory environment that views biometric data as a permanent, and private, asset.
With reporting from The Next Web.
Source · The Next Web
