Amazon Bedrock Guardrails enhances generative AI application safety with new capabilities
AWS News Blog
AWS has announced new capabilities for Amazon Bedrock Guardrails to enhance the safety and governance of generative AI applications:
- Multimodal toxicity detection with up to 88% accuracy for detecting harmful image and text content
- Enhanced PII detection and masking for input prompts
- IAM policy-based enforcement to mandate guardrail policies across AI interactions
- Selective guardrail policy application to optimize performance
- New monitor mode for analyzing guardrail effectiveness before deployment
These features help organizations implement responsible AI practices, maintain consistent content filtering, and protect user privacy across generative AI applications.
The AWS News Feed is currently looking for gold sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.
Related articles
Apr 8
2025
2025
Amazon Bedrock Guardrails announces new capabilities to safely build generative AI applications
Jan 15
2026
2026
Safeguard generative AI applications with Amazon Bedrock Guardrails
Mar 2
2026
2026
Build safe generative AI applications like a Pro: Best Practices with Amazon Bedrock Guardrails
Oct 16
2025
2025
Protect your generative AI applications against encoding-based attacks with Amazon Bedrock Guardrails
The AWS News Feed is currently looking for silver sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.