Google: Gemini AI stops malicious ads

Cybercriminals are increasingly turning to the capabilities of Generative AI, creating a new wave of ad fraud with unprecedented speed and scale. The automation offered by AI allows for the mass production of misleading ads, which are increasingly difficult to detect with traditional methods. The result is an ever-evolving threat landscape that tests the limits of existing security mechanisms. In the face of this new reality, Google is attempting to shield the advertising ecosystem by integrating advanced AI Gemini models into its infrastructure. This strategy marks a transition from passive filters to dynamic, automated defense systems that operate in real time.

Google's response by leveraging Gemini AI

According to the 2025 Ad Security Report, the integration of Gemini models significantly enhanced malicious content detection capabilities. The new systems were able to block the vast majority of malicious ads before they were displayed to users, drastically reducing exposure to fraud.

See also: How targeted advertising works in depth

This transition is not just a technological upgrade, but a total change in philosophy for ad security. Google is now investing in proactive mechanisms that detect threats before they occur, rather than reacting after the fact.

From keywords to behavioral analysis

Traditional keyword-based filters are proving inadequate against modern spoofing techniques. Attackers use sophisticated obfuscation methods, easily bypassing simple content checks.

Gemini models change the way data is analyzed, evaluating vast amounts of information in real time. Instead of focusing solely on the content of an ad, they analyze account behavior, campaign patterns, and dozens of other signals that reveal the true intent behind each action. This makes detection more meaningful and less superficial.

Massive interventions and impressive results

The application of these technologies had immediate results at a global level:

• Blocking or removing over 8.3 billion malicious ads worldwide
• 24.9 million advertiser accounts suspended for serious policy violations
• Blocking 602 million ads specifically linked to fraudulent scams
• Deactivating 4 million accounts directly related to active fraud campaigns

See also: Google Gemini: Improving answers to mental health questions

Defense against AI-generated fraud

The same technology used by attackers is now being used on the defensive. Google is implementing Gemini to detect and block fraud created by automated tools before it enters the advertising ecosystem.

The ability to assess ads at the time of submission is a critical advantage, as it prevents threats from spreading from the source. The company plans to expand this approach to other forms of advertising, further strengthening defenses.

Reducing errors and protecting legitimate businesses

One of the key issues with such systems is the risk of false blocking. However, Gemini's advanced analysis allows for more accurate discrimination between legitimate and malicious ads.

This improvement has resulted in a significant reduction in unfair account suspensions, allowing legitimate businesses to continue operating without disruption. Combined with stricter verification processes, Google is working to create a safer and more trustworthy environment for everyone.

See also: Google Home update: Better command understanding from Gemini

The next day in digital ad security

The battle between attackers and defense systems is expected to intensify in the coming years as artificial intelligence advances on both sides. Google's approach suggests that the future of security will be based on intelligent, automated systems that constantly learn and adapt.

In this environment, user and business trust will depend on platforms’ ability to identify threats early without stifling innovation. This balance will be a key factor in the sustainability of digital advertising in the new era.