AnalysisPublished 10 Jun 2026

The guide on blocking ChatGPT, Gemini, Claude, and other AI tools at work | Kaspersky official blog

Source Kaspersky Security Blog

Description

How to shut down unauthorized AI tools and block user access across a corporate network.

Impact

The impact described is the risk of data leaks and security breaches resulting from unauthorized use of AI tools in corporate environments. Employees may inadvertently expose sensitive company data to public AI services or install unapproved AI applications that could be insecure or exploited. The guide does not describe a software vulnerability or direct exploitation but focuses on the operational risk of shadow AI usage leading to potential data exposure and compliance issues.

Defensive guidance

This is a security best practices guide rather than a vulnerability with a patch. Organizations should implement the recommended detection and blocking strategies using NGFWs, web filters, DNS rerouting, EDR/EPP tools, browser policies, application allowlisting, and mobile device management (MDM) to restrict unauthorized AI tools. Blocking known AI service domains and applications, enforcing strict extension policies, and auditing meeting participants for AI bots are advised. The guide also recommends providing approved AI alternatives to reduce users seeking unauthorized services.

Summary

The Kaspersky blog post provides detailed technical guidance on how organizations can detect and block unauthorized AI tools and services on corporate networks and devices. It covers network-level detection via next-generation firewalls (NGFW) and web filters monitoring traffic to known AI service domains, endpoint detection using EDR/EPP tools to scan for AI-related applications and browser extensions, and policy enforcement through browser and application controls. The guide includes specific instructions for blocking popular AI tools such as ChatGPT, Claude, Perplexity AI, DeepSeek, Slack AI, Zoom AI Companion, Grammarly, AI meeting assistants, AI code editors, local AI tools, and autonomous AI agents. It highlights the security risks posed by unauthorized AI use, including data leaks and potential cyberattacks on local AI applications. The document is advisory and does not describe a software vulnerability or exploit.

References

kaspersky.com/blog/how-to-detect-disable-shadow-ai-in-enterprise/55952/ (opens in a new tab)

Community reviews

Loading reviews…

Description

AI analysisMachine-generated, updated 10 Jun 2026, 16:47 UTC

References

Community reviews