The AI company received the photos from OkCupid in 2014 without users' knowledge, in a data transfer that breached OkCupid's own privacy policy.
The FTC's late March settlement with OkCupid and Match Group brought no financial penalties. Clarifai was not accused of wrongdoing.
Clarifai, a Delaware-based facial-recognition AI company, has confirmed that it deleted approximately three million OkCupid user photos and the facial-recognition models trained on those images, following the US Federal Trade Commission's settlement with the dating site over a privacy violation that dates back to 2014.
Clarifai certified its deletion to the FTC on 7 April 2026, according to a document seen by Reuters, and told the office of US Representative Lori Trahan on 16 April that it had deleted any models trained on the data and had not shared it with third parties.
The underlying incident began over a decade ago. OkCupid's founders were investors in Clarifai, and Clarifai's founder Matthew Zeiler contacted OkCupid co-founder Maxwell Krohn in 2014 to request access to its data.
The 💜 of EU tech
The latest rumblings from the EU tech scene, a story from our wise ol' founder Boris, and some questionable AI art. It's free, every week, in your inbox. Sign up now!
“We're collecting data now and just realized that OKCupid must have a HUGE amount of awesome data for this,” Zeiler wrote, according to court documents cited by Reuters.
OkCupid handed over nearly three million user photos, along with location and demographic data, without any formal agreement, without placing restrictions on how the data could be used, and without notifying users or allowing them to opt out.
OkCupid's privacy policy at the time explicitly stated it would not share personal data with parties outside a defined set of business relationships. Clarifai was none of those things.
The FTC opened its investigation after a New York Times article in 2019, but the case took years to settle. The proposed consent order, announced on 30 March 2026, prohibits OkCupid and its parent Match Group, which also operates Tinder, from misrepresenting their data practices for 20 years.
It does not include a financial penalty: the FTC does not have authority to issue fines for violations of this type under Section 5 of the FTC Act. Clarifai was not accused of any wrongdoing, having received the data through a request rather than initiating the transfer.
The settlement drew immediate criticism. Representative Trahan, a Democrat from Massachusetts, described Clarifai's deletion confirmation as “a step in the right direction,” but added that “the FTC should have never settled for less in the first place.”
The case is the FTC's first Section 5 privacy enforcement action under Chair Andrew Ferguson, and legal analysts at Venable note that, unlike many prior FTC privacy consent orders, this one imposes no ongoing compliance programme requirements or affirmative notification obligations on the companies.
The technical profile of the data makes the absence of penalties more striking. Clarifai's products include facial-recognition systems that can identify individuals and analyse age, race, and gender from images.
The company has contracted with the US military and received investment from Nvidia. Facial-recognition models trained on intimate social data, photos people shared on a dating platform, often expecting privacy, carry different risks from models trained on publicly available images.
Whether those models persist in any form elsewhere in the AI ecosystem, through third-party licensing or derived models, remains unknown: Clarifai did not say how long the models were in operation before deletion.