Approximately 3 million user images from the dating platform OkCupid were utilized for AI training by an external organization, breaching the company’s privacy policy. This violation dates back to 2014, but it was only recently addressed, leading to the deletion of both the photos and the AI model derived from them.
The situation commenced when Clarifai’s founder, Matthew Zeiler, reached out to OkCupid co-founder Maxwell Krohn, seeking access to its data. Zeiler acknowledged the significant potential of OkCupid’s data for their AI projects. OkCupid furnished nearly three million user images, along with location and demographic information, without a formal contract, limitations on data usage, or informing users.
In spite of OkCupid’s privacy policy vowing not to distribute personal information with unrelated third parties, the data was shared informally. The Federal Trade Commission (FTC) initiated an investigation in 2019, which concluded this year. As part of the settlement, OkCupid’s parent company, Match Group, has been prohibited from misrepresenting its data practices for two decades. Nevertheless, the FTC does not possess the capacity to levy monetary penalties for such privacy infringements.