CLARIFAI DELETES 3M OKCUPID PHOTOS AFTER FTC SETTLEMENT

AI DESK■ 1 MIN READ

TUE, APR 21, 2026

■ AI-SUMMARIZED FROM 2 SOURCES ▸ TIMELINE

AI company Clarifai has deleted 3 million OkCupid user photos and facial-recognition models trained on them following an FTC privacy settlement with the dating platform.

Clarifai removed the dataset after regulators concluded OkCupid violated user privacy by sharing photos without consent. The dating app provided the images to Clarifai in 2014 to train facial recognition systems, according to court documents. OkCupid executives had invested in Clarifai, creating a financial connection that influenced the data-sharing arrangement. The FTC settlement requires OkCupid to obtain explicit user permission before sharing personal information with third parties and to delete previously shared data. The case highlights ongoing regulatory scrutiny of how tech companies handle user data for AI training. Federal regulators have increasingly targeted firms that use personal information without explicit consent, particularly for facial recognition and other biometric applications. Clarifai confirmed the deletion in recent statements, though the exact date of the FTC settlement was not specified in available reports.

■ SOURCES

► Techmeme ► TechCrunch

■ SUMMARY WRITTEN BY AI FROM THE LINKS ABOVE

■ MORE FROM THE SECURITY DESK

P384CHROMIUM 148 MATH.TANH CREATES NEW OS FINGERPRINT

A new browser fingerprinting vector has emerged in Chromium 148, where the Math.tanh function produces different results across operating systems. This discrepancy can be exploited to identify a user's underlying OS without explicit permission.

8H AGO— Industry Desk

P373MSPS RETHINK SECURITY AS BREACHES TEST RECOVERY

Kaseya is hosting a webinar on strengthening MSP resilience through SaaS backups and business continuity strategies. The session focuses on how recovery capabilities prove critical when security defenses are breached.

14H AGO— Security Desk

P370REDHOOK MALWARE EXPLOITS WIRELESS ADB FOR SHELL ACCESS

A new variant of RedHook Android malware abuses Wireless ADB (Android Wireless Debugging) to gain shell-level privileges without requiring a computer connection. This represents a significant escalation in the malware's capabilities.

15H AGO— Security Desk

P358FAKE NEWS ARTICLES LURE VICTIMS TO INVESTMENT SCAMS

Fraudsters are creating convincing counterfeit news articles impersonating major publishers like the Guardian to direct social media users to bogus investment sites. The fake stories feature fabricated celebrity endorsements and financial narratives designed to establish credibility.

19H AGO— Industry Desk

◄ BACK TO NEWS

CLARIFAI DELETES 3M OKCUPID PHOTOS AFTER FTC SETTLEMENT

■ MORE FROM THE SECURITY DESK

■ SUBSCRIBE TO THE DAILY BRIEF