Google says AI blocking billions of fraudulent ads before users even see them

Gemini AI credited with intercepting 99% of harmful content before it goes online

Life

Image: Allen Boguslavsky via Pexels

While Internet users are bombarded with ads on a daily basis, bad actors are increasingly trying to use them to spread misleading or dangerous messages. Google is therefore deploying its safety teams and its Gemini AI model to stay one step ahead of these scammers. That approach is paying off: in 2025, the tech giant blocked 8.3 billion ads worldwide. Strikingly, 99% of these violations were intercepted before any user ever got to see the harmful content.

The problems with rogue ads occur everywhere. In the European Union, Google took firm action last year. There, the company removed 1.6 billion ads and suspended 2 million advertiser accounts. This emerges from the tech giant’s annual Ads Safety Report. Most violations in Europe related to abuse of the advertising network and trademark infringements. In addition, the systems tackled misleading ads and breaches involving personalised advertising.

Bad actors are increasingly turning to generative AI themselves to evade detection, resulting in a digital game of cat and mouse. Google uses Gemini to analyse hundreds of billions of signals at high speed. The model looks at the age of an account and flags unusual campaign patterns. Last year, 602 million removed ads and four million suspended accounts were directly linked to scams. Gemini also helps process user reports more efficiently. As a result, Google was able to follow up on four times as many complaints as the year before.

It is crucial that legitimate companies are not hampered by these strict checks. Gemini is far better at understanding the intent behind an ad than older systems that relied purely on keywords. This allows the AI to draw a much sharper line between a genuine offer and a scam. In 2025, this resulted in 80% fewer wrongful suspensions. Trustworthy companies can therefore continue their campaigns without disruption, while Google weeds out the rotten apples from its network.

Business AM

Read More: AI Artificial Intelligence Google