Everalbum's Ever photo app used facial recognition technology on millions of user photos without proper consent to develop commercial facial recognition software sold to enterprise customers, while also failing to delete photos of users who deactivated their accounts.
Everalbum, Inc., a California-based company, operated a photo storage app called Ever that allowed users to upload and store photos and videos. In February 2017, the company launched a 'Friends' feature that used facial recognition technology to group photos by faces and enable tagging, with facial recognition enabled by default for all mobile users. Between July 2018 and April 2019, while Everalbum claimed it would not apply facial recognition without user consent, it automatically activated the feature for users outside Illinois, Texas, Washington, and the EU, where it could not be turned off until April 2019. Between September 2017 and August 2019, Everalbum extracted millions of facial images from user photos and combined them with publicly available datasets to create four datasets for developing facial recognition technology. This technology was used not only for the Friends feature but also to develop facial recognition services sold to enterprise customers, including potential law enforcement and military clients. The company also promised to delete photos and videos of users who deactivated accounts but failed to do so until at least October 2019, retaining them indefinitely. The FTC settlement requires Everalbum to delete user photos from deactivated accounts, face embeddings from users who did not consent, and any facial recognition models developed using user content. Everalbum shut down the Ever app in August 2020 and rebranded as Paravision AI.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that memorize and leak sensitive personal data or infer private information about individuals without their consent. Unexpected or unauthorized sharing of data and information can compromise user expectation of privacy, assist identity theft, or cause loss of confidential intellectual property.
Human
Due to a decision or action made by humans
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed