Google Chrome adds Gemini AI scam detection as OpenAI mounts in-browser challenge
Chrome 156 fixes over 300 vulnerabilities while adding server-side threat detection and media analysis to counter a renewed push from OpenAI.

Google has upgraded its Chrome browser with server-side Gemini artificial intelligence to block deceptive scams and summarise video content, launching the defensive tools just as OpenAI rolls out an aggressive update to its own browser extension.
The new security capabilities deploy Gemini AI models to analyse web page content, detect fraudulent tech-support warnings, and stop ClickFix attacks. These attacks trick users into copying malicious commands into command-line utilities like PowerShell. To counter them, Chrome strips personally identifiable information from suspicious clipboard payloads before running them through Google Safe Browsing.
Alongside security enhancements, Chrome version 156 introduces a media understanding feature that allows users to ask Gemini about video or audio playing in their active tab. The update also addresses 353 security vulnerabilities, including six critical memory errors, with Google paying $50,000 in bounties to external researchers who reported flaws.
However, Google's efforts to establish Gemini as the default helper inside its browser face direct competition. OpenAI has updated its official ChatGPT extension for Chrome, introducing a sidebar that reads across open tabs and automates tasks on websites through ChatGPT Work.
| Feature or Capability | Google Gemini in Chrome | OpenAI ChatGPT Extension |
|---|---|---|
| Integration type | Native browser features | Web Store extension |
| Threat detection | Server-side scam & ClickFix analysis | Not offered |
| Media analysis | Summarises active video and audio | Tab text and page context |
| Cross-tab context | Integrated search and settings | Reads across multiple open tabs |
| Web automation | Not disclosed | Executes tasks via ChatGPT Work |
How Gemini detects scam websites in Chrome
Google said its upgraded scam detection moves beyond on-device checks to evaluate page content on Google servers whenever high-confidence scam indicators appear. The system evaluates brand representation and assigns a scam-likelihood score, which Safe Browsing combines with domain reputation metrics before issuing a warning.
Text sent for server-side evaluation is processed ephemerally and is not logged, retained, or used to train AI models, Google stated. Chrome has also introduced a lower-confidence suspicious category with distinct visual warnings, reserving full red warning screens for confirmed malicious destinations. Access to these protections requires users to turn on Enhanced Protection in Chrome settings.
ChatGPT challenges Gemini on Google's home turf
OpenAI's latest extension update allows ChatGPT to assist users without requiring them to switch between browser tabs or navigate away from active pages. The assistant can compare articles, organise tabs, and extract details from multiple open websites.
Through its ChatGPT Work tier, the extension goes a step further by taking actions directly on web pages, such as filling out forms or gathering research across multiple open tabs. While Google holds a structural advantage by bundling Gemini natively into Chrome, OpenAI's update turns the browser into a direct arena for AI assistant adoption.
Key numbers
| What | Figure | Source |
|---|---|---|
| Security vulnerabilities fixed | 353 | Chrome Releases blog post by Srinivas Sista |
| Critical vulnerabilities fixed | 6 | |
| Total bug bounties awarded | $50,000 | |
| Flaws found by external researchers | 87 |
In this story
| Name | Role in the story |
|---|---|
| Gemini | Google's AI assistant embedded into Chrome. |
| OpenAI | Developer of ChatGPT updating its Chrome browser extension. |
| Developer of Chrome and Gemini AI tools. | |
| ChatGPT | AI assistant expanding capabilities inside Chrome. |
| Chrome | Web browser receiving Gemini AI scam detection and security updates. |



