Clarifai Confirms Deletion of 3 Million OkCupid User Photos and Facial-Recognition Models

Clarifai Confirms Deletion of 3 Million OkCupid User Photos and Facial-Recognition Models

2 Min Read

Clarifai, a Delaware-based facial-recognition AI company, confirmed the deletion of three million OkCupid user photos and the facial-recognition models trained on them following the US Federal Trade Commission’s settlement with the dating site over a 2014 privacy violation. Clarifai certified the deletion to the FTC on 7 April 2026, and informed US Representative Lori Trahan on 16 April that models trained on the data were deleted and not shared with third parties.

The incident began over a decade ago when OkCupid’s founders, also investors in Clarifai, were approached by Clarifai’s founder, Matthew Zeiler, who requested access to OkCupid data. In 2014, OkCupid provided nearly three million user photos with location and demographic data without any formal agreement, user notification, or opt-out options.

OkCupid’s privacy policy at the time prohibited sharing personal data outside specified business relationships, of which Clarifai was not part. The FTC investigation began after a 2019 New York Times article, leading to a consent order on 30 March 2026, barring OkCupid and parent Match Group from misrepresenting data practices for 20 years. No financial penalty was imposed, as the FTC lacks authority for fines under Section 5 of the FTC Act. Clarifai, having received the data via request, was not accused of wrongdoing.

The settlement faced criticism, with Representative Trahan from Massachusetts acknowledging Clarifai’s deletion as positive but criticizing the FTC’s settlement as insufficient. This case, the FTC’s first Section 5 privacy enforcement under Chair Andrew Ferguson, didn’t require compliance programs or notification obligations, unlike past cases.

Clarifai’s facial-recognition systems can identify individuals and analyze age, race, and gender. The company has military contracts and Nvidia investment. Models trained on private social data pose different risks compared to those using public images. Whether these models persist elsewhere in the AI ecosystem is uncertain; Clarifai did not disclose their operational duration before deletion.

You might also like