A bipartisan Senate investigation has laid bare the structural negligence at the heart of America's largest dating platform company. Internal documents from Match Group — owner of Tinder, Hinge, Plenty of Fish, and Match — show that as of late 2025, Tinder had not updated most of its machine learning models designed to detect scammers "in years." The result was predictable: nearly 1.5 million scammer or bad-actor accounts on Tinder and Hinge matched with at least one real user through the third quarter of 2025, and roughly 500,000 of those accounts successfully exchanged contact information with their targets. The human cost is not abstract. Mary, a Plenty of Fish user, lost nearly $100,000 in cryptocurrency payments to a romance scammer who maintained constant contact, knew her location, and fabricated work emergencies and medical crises to extract money. She described the experience as "being in a vortex" — each payment felt disconnected from the last until the total became devastating. When she blocked the scammer, he simply switched phone numbers and resumed contact. She later discovered his photos were being recycled across multiple profiles on the platform. The joint economic committee report, led by Senators Maggie Hassan (D-NH) and Marsha Blackburn (R-TN), places Match Group's negligence within a $1.3 billion annual crisis. Romance scams have become a leading form of financial fraud in the United States, and the report makes clear that the company's defenses degraded precisely as criminal networks became more sophisticated. The outdated ML models on Tinder meant the platform "detected scammers and other bad actors less reliably and struggled to keep pace with emerging threats." Match Group's response follows a familiar corporate playbook: claim credit for improvements while contesting the timeline. Senior VP of trust and safety Yoel Roth said the report "mischaracterizes both the historical scale of our investment and the work already complete," and the company told NBC News that safety features like facial verification were already under development before the investigation began. The report itself acknowledges that Match Group has since introduced facial verification requiring video selfies for new Tinder users and deployed a new AI model to detect scammer profiles. The core tension is one of incentive structure. Match Group profits from user engagement. Scammer accounts that match with real users, exchange messages, and create the illusion of connection generate the same engagement metrics as legitimate ones — potentially better, given how aggressively scammers maintain contact. The company had no commercial incentive to aggressively purge accounts that were, from a metrics standpoint, highly active users. Congressional pressure appears to have been the forcing function. Hassan has also introduced legislation with Republican Senator Rick Scott to establish a central scam-reporting portal — an acknowledgment that the current enforcement landscape is fragmented and that victims like Mary have no clear institutional pathway for recourse. The broader policy question is whether platform self-regulation can ever work when the platform's revenue model is structurally misaligned with user safety. The report describes romance scams as the work of "global criminal enterprises," and Hassan warns that AI is being used to "supercharge" these operations. The implication is stark: the offensive capability of scam networks is scaling faster than the defensive capability of platforms that have financial reasons not to invest aggressively in detection. Without regulatory mandates tied to enforcement, the pattern will repeat.
