Google is experimenting with a new security feature in Chrome Canary designed to leverage artificial intelligence to identify potential online scams. This feature, known as “Client-Side Detection Brand and Intent for Scam Detection” (an official translation hasn’t been provided yet), was recently uncovered and shared by user @Leopeva64 on X (formerly Twitter).
Found in the browser’s experimental Flags menu, this innovation represents a significant step forward in safeguarding users against cyber threats.
How does Chrome’s new AI security feature work?
The new feature employs large language models (LLMs) to analyze websites in real time, helping to identify potential fraud attempts. While specific technical details haven’t been disclosed, it’s likely that the system is being tested in scenarios commonly associated with online scams.
One possible application is detecting fake technical support pages, such as those masquerading as official Microsoft support sites. The LLM could evaluate the page’s content, URL structure, and other elements to determine if it’s legitimate or a phishing attempt. If the system identifies suspicious traits, it alerts users about the potential risk.
This functionality complements Chrome’s existing “Safe Browsing with Enhanced Protection” feature, which also uses AI to guard against malicious websites, risky downloads, and suspicious extensions. However, the new tool seems to focus specifically on detecting scams and fraud, adding an extra layer of security to Chrome’s already robust protection measures.
Using pre-trained data
Google’s new scam detection feature likely leverages vast pre-trained datasets to power its large language model (LLM). These datasets could include malicious URL patterns, typical layouts of fraudulent pages, and common textual cues associated with phishing attempts.
By analyzing the content displayed to users in real-time, the system identifies matches or similarities with known scam patterns. This enables dynamic and adaptive protection, making it more capable of detecting new and evolving threats.
This implementation marks a major step forward in proactive cybersecurity. Instead of relying solely on static blocklists or basic heuristics, AI enables more sophisticated and context-aware analysis of web pages, which could significantly enhance its ability to catch emerging scams.
That said, as with any AI-driven system, false positives or negatives are possible. The tool’s ultimate effectiveness will depend on rigorous testing and user feedback during its experimental phase.
For now, there’s no official timeline for when “Client Side Detection Brand and Intent for Scam Detection” will roll out to the general public. Currently, it remains exclusive to the experimental Chrome Canary version.
