AI Guardrails is a content moderation service that complements data protection and threat protection by:
-
protecting from AI threats like prompt injection and jailbreaking.
-
enforcing content moderation guardrails for responsible AI usage by detecting and protecting from many categories of AI prompts and responses, such as inappropriate/harmful content, hate, bias, weapons as well as copyright or patented data that expose a new risk vector for organizations.
-
leveraging a new AI inspection engine that classifies the prompts and responses by mapping them to MITRE ATLAS and OWASP Top10 LLM frameworks for adversary tactics, techniques, and real-world case studies targeting AI systems.
AI Guardrails is supported in FedRAMP and PBMM tenants.
Supported Applications
Netskope supports protection by scanning prompts and responses for the following popular apps:
-
Anthropic
-
Amazon Bedrock
-
Atlassian Confluence
-
Atlassian Jira
-
Box
-
ChatGPT
-
Citrix ShareFile
-
DeepSeek
-
Egnyte
-
GitHub Copilot
-
Google Gemini
-
Google Gemini Enterprise
-
Google NotebookLM
-
Google NotebookLM Enterprise
-
Grammarly
-
Grok
-
Humata
-
Microsoft 365 Copilot
-
Microsoft Copilot
-
Microsoft Copilot Studio
-
Microsoft Foundry
-
Mistral Vibe
-
Prezent
-
Quillbot
-
Salesforce
-
Twitter
-
YouTube
With more apps coming soon.
Supported Languages
-
East Asian Languages
-
Chinese (Simplified & Traditional)
-
Japanese
-
Korean
-
-
Major European Languages
-
English
-
French
-
Spanish
-
German
-
Italian
-
Portuguese
-
Russian
-
Dutch
-
Polish
-
Ukrainian
-
Bulgarian
-
Romanian
-
Czech
-
Danish
-
Finnish
-
Norwegian
-
Swedish
-
-
Other Languages
-
Arabic
-
Hindi
-
Vietnamese
-
Thai
-
Turkish
-
Hebrew
-
Filipino (Tagalog)
-
Catalan
-
Indonesian
-