AI TraceTrace Foundation, Inc.
Content ModerationConsumer FacingVerified

Reviewed and published by trentmaziarz, March 22, 2026. Discovered and drafted by our automated research pipeline.

In December 2018, Tumblr deployed an AI image-scanning system to automatically flag and hide content showing nudity or sexual activity, as part of a complete ban on adult content. The system was deeply flawed from launch, incorrectly flagging hundreds of thousands of harmless images — including cartoons, patent drawings, pet photos, and even images of then-Vice President Joe Biden — while missing actual prohibited content.

Details

The ban was triggered by Apple removing Tumblr's iOS app on November 16, 2018, after child sexual abuse material (CSAM) slipped past existing detection systems. Tumblr responded by announcing a full adult content ban enforced using deep learning image classifiers combined with human moderator review of appeals. The classifiers produced a catastrophic number of false positives at launch, and Tumblr's own example images of "permitted content" were flagged by its own filter within a day. Traffic fell from roughly 521 million monthly visits in December 2018 to 370 million by February 2019 — a 29% drop — and Tumblr's acquisition price by Automattic in August 2019 was approximately $3 million, down from Yahoo's $1.1 billion purchase price in 2013. In February 2022, Tumblr settled with New York City's Commission on Human Rights after findings that the ban disproportionately affected LGBTQ+ users; the settlement required Tumblr to hire an expert to audit its moderation algorithms for bias and to replace the original classification system.

Products affected

Tumblr

Sources & Evidence

Other practices by Tumblr

Have evidence about Tumblr's AI practices? Submit a report.

Report a Sighting →