Tinder Data Scraping: 40,000 Selfies Used for AI Training
A Tinder user, Stuart Colianni, has scraped 40,000 profile pictures from the dating app to create a facial dataset called “People of Tinder.” Colianni, a user of the machine learning platform Kaggle, released the dataset under a public domain license, making it freely available for download.
The dataset includes photos from both men and women in the Bay Area, and Colianni claims it offers “near unlimited access to create a facial dataset.” He argues that Tinder’s API provides an “extremely efficient way to collect such data” and criticizes other datasets for being too small and restrictive.
However, Colianni’s actions have sparked controversy due to the privacy implications. Many users are unaware that their photos could be used for AI training without their consent. While Tinder’s terms of service grant the company rights to use user content, it’s unclear if this extends to third-party developers scraping data and releasing it publicly.
The dataset has been downloaded over 300 times, raising concerns about potential misuse. Some users have expressed their discomfort with their photos being used for research without their knowledge or permission.
Tinder has issued a statement condemning the scraping, stating that it violates their terms of service and that they are taking appropriate action. The company is investigating the matter further and plans to implement additional measures to prevent API misuse.
