A new feature in the Chrome browser is supposed to warn users when a site is trying to scam them. The feature, which appears to use AI, is currently being tested.
If you enable the feature, Chrome will automatically scan various web pages for possible scams. It uses a large language model (LLM) for this on your device itself. We say it is because the feature is not usable for the general public yet. A researcher on X/Twitter discovered them in Chrome Canary, a browser’s beta version. According to the description, the LLM looks for the brand behind the web page and tries to determine its intentions. That way, it could also identify possible scams.
Interestingly, the LLM appears to be on the user’s device, and from there, it analyses your browser traffic. This way, your surfing data would not be sent to a server somewhere in the cloud, which should improve privacy.
Edge
Scanning for scams is a form of security that browsers are increasingly trying to make a difference with. Earlier this month, Microsoft announced a similar feature for its Edge browser. That tech giant also wants to use an LLM to pick up so-called ‘scareware’. That is the type of scam where a user gets (fake) pop-up windows with security warnings. If Edge finds that kind of malware on a site, users can close or ignore the warning.