---
sourceDocument: Yokohama Enable AI
sourceDocumentLink: https://servicenow-prod.fluidtopics.net/r/yokohama/intelligent-experiences

 Release :

    - yokohama

ft:locale :

    - en-US

ft:publication_title :

    - Yokohama Enable AI

ft:clusterId :

    - platai

bundleId :

    - platai

workflow :

    - Platform


---

# Configure prompt injection attack protection

# Configure prompt injection attack protection {#ariaid-title1}

Release version: Yokohama  
Updated July 31, 2025  
![](https://www.servicenow.com/docs/portal-asset/ico-clock) 1 minute to read  
Activate or deactivate prompt injection attack detection to protect all generative AI applications and AI-generated text and conversations.

## Before you begin

Role required: sn_generative_ai.nsa_admin

## About this task

Prompt injection attacks are a type of cybersecurity attack where someone tries to override the initial instructions of an LLM to cause unintended behaviors. AI Guardian detect and log these prompt injection attack attempts across all generative AI applications and features. You can also configure the prompt injection detection guardrail to block the
AI-generated response when an attack is detected in addition to logging it.

You can export logs for review. For more information, see [Export Now Assist Guardian logs](https://servicenow-prod.fluidtopics.net/n_5MubOPi1HhhBTXleT4EQ "Export logs from AI Guardian to get insights into how often different guardrails are being detected and used.").

## Procedure

1. Navigate to AllNow Assist AdminSettings.
2. In the side panel, go to Now Assist GuardianPrompt Injection.
3. Select the toggle to activate or deactivate prompt injection attack detection.
4. **Optional:** Under Detection impact, select the options icon (![Options icon.]()) and then select Edit to change how detected attacks are handled.  
   You can choose whether prompt injection attacks are blocked as well as logged.

## Result

Prompt injection detection is configured on your instance. When enabled, you see a standard error message when an attack is detected.

*[\>]: and then