Google to ban Chrome extensions that jailbreak AI

Google is closing the door on Chrome extensions built to help users dodge AI safety limits, part of a policy overhaul announced for the Chrome Web Store this week.

The centrepiece of the update is a new addition to Chrome's Malicious and Prohibited Products Policy that explicitly bans extensions designed to circumvent safety guardrails, usage restrictions, or other protective measures built into AI-powered services.

Until now, tools marketed as ways to unlock unrestricted chatbot responses or bypass usage caps had no clear policy standing in the way, according to Google's developer blog.

In addition to AI, Google is making changes to its Limited Use Policy, which means that data collected by extensions will have to be absolutely necessary for achieving a single purpose. It will not be allowed for developers to collect additional data, as it may become relevant for some other feature in the future.

An additional amendment will be made to the Disclosure Requirements Policy, meaning that developers will have to disclose their data collection and inform users about it even if data collection is essential for achieving an extension's purpose and has been changed after installation.

Google also made updates to its policy on regulated goods and services, which now specifically prohibits any extension that enables real-money transactions based on predictions, thus effectively shutting down a sector which was increasingly prominent in the Chrome Web Store but lacked any specific policy guidelines.

Google noted that enforcement starts August 1, 2026, giving developers about a month to get their current extensions up to code. Extensions which do not comply with the new rules will face enforcement actions and may be taken down from the Chrome Web Store.