A new scam is making the rounds online, and this time it’s masquerading as the ChatGPT Atlas browser. Security researchers are warning users to be careful after spotting a convincing fake version promoted through search ads—one that looks legitimate at first glance but is designed to steal your account data.

The warning comes from research firm Fable. The fake browser isn’t using any sophisticated exploits. Instead, it relies on tricking people who assume anything that looks polished and appears at the top of search results must be safe.

The setup is simple. A search ad appears promising a download of “ChatGPT Atlas,” complete with branding that matches the real thing. Clicking on the link takes you to a website that looks nearly identical to the official one. The layout, the wording, even the design cues are copied almost perfectly. But there’s one major giveaway: the site is actually hosted on Google Sites. According to Fable, scammers often clone legitimate pages using tools like v0.dev, then deploy them on Google’s hosting platform to give the illusion of trustworthiness—especially for users who associate “Google” with safety.

Things get shady the moment you try to download the app. Instead of offering a normal installer file, the site tells users to paste a command into their terminal. Anyone even slightly tech-savvy would recognize this as a massive red flag. But for people who aren’t familiar with the feature, it might look like a harmless setup step.

The command itself is disguised as a base64-encoded string that gets decoded and executed through curl and bash. Once it runs, the malware pops up a request asking for your administrator password. If the user enters it, the trojan gains full sudo privileges and immediately installs a second-stage payload. From that point on, the attacker can scrape stored browser passwords, account logins, and other sensitive data.

This tactic is essentially a twist on the known “ClickFix” attack method. Nothing here is technically advanced—it’s almost shockingly simple. But it works because it leans heavily on social engineering: people trust familiar brands, trust search ads, and trust anything that looks clean and official. And that’s exactly what the attackers are banking on.

The takeaway is clear: if a website ever tells you to paste a terminal command to install something, stop right there. Always double-check the domain, and only download apps from verified sources.

Don’t miss a thing! Join our Telegram community for instant updates and grab our free daily newsletter for the best tech stories!

For more daily updates, please visit our News Section.

(Via)

The post Beware the fake ChatGPT Atlas browser stealing account logins and saved passwords appeared first on Gizmochina.

OpenAI has launched ChatGPT Atlas, a new AI-powered browser that integrates its popular chatbot directly into the browsing experience. The company made the browser available on macOS this week, with support for Windows, iOS, and Android expected soon. Atlas is free to download and supports all ChatGPT account tiers, including Free, Plus, Pro, Business, and Enterprise.

ChatGPT Atlas replaces the traditional address bar with a chat-based interface, allowing direct interaction with AI features across any webpage. It eliminates the need to switch tabs or copy text, providing assistance in real time. The built-in sidebar enables ChatGPT to summarize articles, answer queries, and assist in completing tasks using the live content of the page.

OpenAI also introduces Agent Mode, a premium feature for Plus and Business subscribers. It enables ChatGPT to handle multi-step actions like online shopping, topic research, and automating routine workflows. It supports contextual assistance, allowing users to highlight text in emails, documents, or invites and receive help directly within the same tab.

The browser includes memory features that let ChatGPT retain browsing context, allowing it to resume tasks such as job searches, list creation, or content summarization. It also offers privacy controls that enable users to disable memory, switch to incognito mode, and limit visibility on specific websites.

Atlas allows data import from Google Chrome, including bookmarks, passwords, and browsing history. It can be set as the default browser through the settings menu, and new users receive higher rate limits during the first week.

OpenAI said the browser is part of its broader effort to reshape how users interact with the web. The launch comes amid growing competition from Google’s Gemini-powered Chrome and other AI-first browsers like Perplexity’s Comet.

According to OpenAI, ChatGPT now serves over 800 million weekly active users. The company aims to expand that number further by integrating AI tools directly into core digital workflows.

For more daily updates, please visit our News Section.

Stay ahead in tech! Join our Telegram community and sign up for our daily newsletter of top stories!

(Source)

The post OpenAI just launched a browser and Google should be worried appeared first on Gizmochina.