Google's own AI would reject this ad

A fake iPhone storage alert kept running on YouTube after repeated reports, and Gemini flagged it in seconds.

Sep 13, 2026

7 cards · 1m 30s · tap to begin

From · · 1m 30s

Google's own AI would reject this ad

A fake iPhone storage alert kept running on YouTube after repeated reports, and Gemini flagged it in seconds.

In brief

A fake iPhone storage alert kept running on YouTube after repeated reports, and Gemini flagged it in seconds. Google already owns a model that catches deceptive ads instantly. The bottleneck is a review process that never asks it. Originally reported by atomic14.

The ad that caught him

The ad that caught him

A YouTube app ad mimicked an iPhone storage warning. He was genuinely low on space, lost concentration for a second, and tapped it.

Reported, and reported again

Well dodgy

He flagged the ad, got a brush-off, flagged it a second time, same answer. Other people reported it too and got the identical reply.

“We found that the ad doesn't go against Google's policies”

Malice or just overload

Ads like this click well, so removal costs money. The kinder read is that human reviewers simply cannot look at everything properly.

“Never attribute to malice that which is adequately explained by stupidity”

It fakes an iOS alert

It

Gemini's verdict: the banner copies iOS typography, container styling and mock Yes/No buttons, so users think their phone is warning them.

Buttons that do nothing

The Yes and No options are static pictures. Tapping anywhere on the banner sends you to the app store listing, not to a real system control.

Fear as the sales pitch

"If you don't free up space soon, some features may not work properly" invents a failure on your device to panic you into installing.

“deceptive claims or scare tactics to induce panic”

Seconds versus two approvals

Gemini disapproved the creative and recommended a policy strike. Google's human process waved the same ad through twice.

“Come on, meatbags - use some of the amazing AI tools you have access to.”

What to remember

Google already owns a model that catches deceptive ads instantly. The bottleneck is a review process that never asks it.

Read the original on atomic14

React

Sign in to react and comment.

Comments (0)

Life is short. Keep it sweet. Respect others' opinions and be kind!

    Recommended next

    More decks on advertising and related topics.