ThreatLens Shadow AI Monitor — Privacy Policy
Last updated: 1 November 2026
ThreatLens Shadow AI Monitor (“the extension”) is an enterprise browser extension that applies an organization’s AI data-protection policy in the browser. It is installed by employees of organizations that use ThreatLens, usually pushed by their IT team, and works only once it is connected to that organization’s ThreatLens environment.
The extension’s single purpose
To apply your organization’s AI data-protection policy to known AI services (for example ChatGPT, Claude, Gemini and Perplexity, on the product’s recognized list plus any AI services your organization adds): check what you send to them against that policy before it is sent, and report use of those services.
What the extension reads, and only on known AI services
- Visits: the URL and time of each visit to a known AI service.
- Prompts: the text of a message when you send it to a known AI service.
- Pastes and files: text you paste, and files you attach, drag or paste into a known AI service, at the moment you do so. Files over 5 MB are not read; your organization’s policy decides whether they may be sent.
- Upload requests: the address, method and result of upload requests to known AI services, to record whether a file was sent or stopped. Never their contents.
It also stores, in your browser, the connection your organization configures: its ThreatLens address, the enrollment token issued to your account, and your organization’s policy settings.
What the extension does NOT read
- Anything on websites that are not known AI services.
- Keystrokes as you type: a prompt is read only when you send it.
- Your clipboard: only what you paste into a known AI service, when you paste it.
- Cookies, passwords or browser credentials. Screenshots.
Permissions the extension uses
- Access to the known AI services (to run the checks and detect visits), ChatGPT’s file-upload address, and your organization’s ThreatLens environment. No access to all websites.
webNavigation: to detect visits to known AI services.webRequest: to see whether an upload to a known AI service was sent or stopped, never its contents.declarativeNetRequest: to stop uploads, or a prompt the policy check didn’t approve, where your organization’s policy requires it.scripting: to run the same checks on AI services your organization adds, and to show the “upload blocked” notice.storageandalarms: to keep the connection and policy settings in your browser and send queued reports.
How the data is used
What the extension reads is sent over HTTPS, with your per-user token, to your own organization’s ThreatLens environment, which checks it against your organization’s policy and answers allow, warn, mask or block. The extension then lets the content through, masks the sensitive part, warns you, or stops it. Your organization’s ThreatLens records each decision (the outcome, the kind of data found, and fingerprints of the content); it keeps the full text or file only if your organization has turned on content retention. Visits feed your organization’s report of AI-service use. ThreatLens does not sell this data and does not use it for advertising, creditworthiness or lending.
Who controls the data
The organization whose ThreatLens environment the extension connects to is the data controller and governs retention and access under its own policies. Where that organization uses ThreatLens as a hosted service, ThreatLens processes the data on its behalf as a processor; where it hosts ThreatLens itself, the data goes straight to that environment and is not processed by ThreatLens. Your connection is revocable from the ThreatLens console, or by removing the extension where your organization allows it.
Retention
Decisions and visits are kept in your organization’s ThreatLens environment for the retention it configures. What the extension stores in your browser is cleared when you disconnect or the extension is removed.
Contact
Questions about this extension: privacy@thethreatlens.com. Questions about how your employer uses the data: your organization’s ThreatLens administrator.