Google Is Working on AI-Powered Scam Detection for Chrome
Scams are everywhere right now. It's never been more important to stay vigilant, whether you're checking your email, browsing the web, or answering a phone call. But it's not all on you: As it happens, Google might be using artificial intelligence to help prevent you from being a victim of an scam—at least while using Chrome.
As spotted by researcher Leopeva64, Google appears to be testing a new scam detection feature in Chrome. Of course, it's 2024, so that means the feature uses AI. The feature's name, "Client Side Detection Brand and Intent for Scam Detection," doesn't alert you to that fact, but you see it in the description: "Enables on devices LLM output on pages to inquire for brand and intent of the page." LLM stands for large language model, which powers many of the AI-generated programs and services you use, and it appears the feature uses that AI model to look out for sites that appear to be scammy.
To further clarify the feature's intent, Leopeva described it to Gemini, which broke down the feature's explanation and distilled it into the following: Based on the bot's analysis, the feature runs an LLM on your device to look for two things: misrepresentation of brands, such as when scammers copy real brands to trick you into accessing their sites; and suspicious intent, which looks for key signs of scams and phishing attempts. You always need to take AI-generated answers with a grain of salt, but, to my eye, that seems like an apt interpretation.
This option is live in the latest version of Chrome Canary, the version of Chrome that Google uses to test new flags (experimental features) on. It's not clear if it works yet, but you can enable it. First, launch Chrome Canary, then head to chrome://flags. From here, search for "client-side-detection-brand-and-page-intent" (without the quotes, of course), then click the "Default" drop-down menu and choose "Enabled." Click "Relaunch," and you're set.
Chrome isn't the only browser getting in on AI scam prevention. Last week, Leopeva64 spotted a similar feature in Microsoft Edge, a "scareware blocker." Unlike Google's more cryptic description, this one quite literally says "Allow Microsoft to use AI to detect potential tech scams."