A US Senate investigation says Tinder had not updated most of its AI models for detecting scammers and other bad actors for years, as scammer profiles drew sharply more views. Match Group disputes the report’s characterisation of its safety work and says it has made major advances.
Watch Desk analysis
What happened
The investigation, led by Democratic Senator Maggie Hassan, examined how Match Group detects and removes scam accounts on Tinder and Hinge. Mashable reports that Tinder had not updated most of its detection models for years as of the second half of 2025. Match’s trust and safety chief, Yoel Roth, told committee staff that the company maintained its machine-learning models and retrained them every year or two.
Figures Match provided to the committee showed that the median scammer profile on Tinder received more than 700% more views between May 2024 and December 2025. In the first nine months of 2025, nearly 1.5 million accounts identified as scammers or other bad actors on Tinder and Hinge in the US matched with at least one user; almost 500,000 exchanged contact information with someone. These are figures reported from Match’s submissions, not independently audited counts.
Match says its Face Check verification process reduced US Tinder users’ interactions with spam, bots and scammers by more than 70% within a month of its nationwide rollout, with the reduction sustained for at least seven months. The company also introduced a new scam-detection model in the first half of 2026, after continued follow-up from the committee. Read Mashable’s report.
Why it matters
The report puts a practical question behind the AI label: how often are detection systems refreshed as scammers change tactics? A model can be part of the defence, but its age, maintenance and results matter to people using dating apps, not just to a company’s product description.
There is evidence of both a problem and a response here. The reported rise in profile views is concerning, while Match’s figures for Face Check suggest the company says some interventions have reduced unwanted interactions. Neither set of numbers settles the broader question of how well the systems work across different scams and users.
Our read
This is a meaningful, specific account of AI being used to address fraud, not a generic claim that an algorithm is involved. Match’s response belongs in the story too: it says the investigation mischaracterises its work and points to safety improvements. The useful next step is evidence that lets people compare the company’s claims with scammer activity and user outcomes. “We retrain the models” is a start; it is not the same as showing what changed for users.
What to watch
- Whether the committee publishes the underlying documents and its full findings.
- How Match measures Face Check’s reported reduction, including its limits and false positives.
- Whether the new detection model produces measurable results as scammers adapt.
Discussion spark: What should dating apps have to disclose to show their scam-detection systems are keeping pace: model update schedules, independent performance tests, or outcomes for users?
Sources and evidence
- Source update (13 October 2026, 18:57 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.