← Cases
United StatesTechnology and telecom2024 – ongoing

Bumble

Deception Detector for spam, scam and fake profiles

Bumble's Deception Detector blocked 95% of flagged spam and scam accounts automatically in testing

WorkingCompany's own claim

Problem

Fake profiles, spam and scams hurt trust on dating apps. Bumble's own survey found 46% of women surveyed were anxious about whether matches were who they claimed to be.

Approach

Bumble built Deception Detector, an AI model that identifies spam, scam and fake profiles and acts on them before members see them. Human moderators work alongside the model. The tool covers Bumble, Badoo and Bumble For Friends.

Results

95%share of accounts identified as spam or scam that the tool blocked automatically in testing, per BumbleCompany's own claim[1]
45%fall in member reports of spam, scams and fake accounts within the first two months of testing, per BumbleCompany's own claim[1]

Takeaway

The model handled most blocking of flagged accounts, and human moderators stayed in the loop. Both figures are Bumble's own testing results.

Categories

Business function
Production and quality
Evidence
Company's own claim

Sources

Source summaries

[1]TechCrunch5 Feb 2024

TechCrunch reported the launch of Deception Detector, an AI feature that identifies spam, scams and fake profiles and works with Bumble's human moderators. It repeated Bumble's figures: 95% of flagged spam and scam accounts blocked automatically in testing, and a 45% fall in user reports within two months of testing.

Full details are in the original sources.

Added 2 Nov 2026 · Updated 2 Nov 2026
Open case page ↗