By Sama Marwan,
Google has announced the deployment of a version of its Gemini artificial intelligence model that runs directly on users’ devices to detect a specific type of cyber fraud known as tech support scams. This technology analyzes page behavior upon opening and alerts users if signs of fraudulent activity are detected. This move is part of Google’s ongoing efforts to enhance user protection across its platforms, including Chrome, Search, and the Android operating system.
The company is leveraging recent advances in AI to combat growing cyber threats, especially as cybercriminals can now generate fake content in massive volumes and with convincing techniques—leading to an unprecedented rise in scams. According to the Global Anti-Scam Alliance, consumer losses from these criminal activities exceeded $1 trillion last year.
Firoozeh Barak, Engineering Director at Google Search, stated that fighting cyber fraud is a constant race against evolving tactics, with scammers always seeking ways to bypass the latest security technologies. She emphasized that the key issue now is not merely the availability of tools, but whether they can be used more effectively than the attackers.
Recent AI updates have significantly improved systems’ ability to understand natural language and detect patterns, enabling Google to identify scam techniques more accurately and quickly than before. The company explained that the “Enhanced Protection” feature in the desktop version of Chrome now uses the Gemini Nano model locally, without needing server communication—ensuring faster response times and better privacy protection for users.
This technology detects scammer tactics such as hiding real content from inspection tools while showing different pages to users. Google has also started alerting Android users when they receive notifications from suspicious websites, with an option to automatically unsubscribe. Additionally, Google is expanding the use of AI to block fraudulent search results across various devices.
Google confirmed that the number of scam pages blocked using AI has increased 20-fold compared to previous years. It also succeeded in reducing scams involving fake customer service numbers by up to 80% in certain sectors, such as aviation.
Meanwhile, other companies are also developing similar solutions. UK telecom provider O2 launched an AI assistant named Daisy to deal with phone scammers. Microsoft developed a tool to analyze calls and detect suspicious activity. The U.S. Department of the Treasury reported that AI helped recover $1 billion from check fraud alone in 2024.
