Automated reasoning checks now available in Amazon Bedrock Guardrails

AWS launches Automated Reasoning checks in Amazon Bedrock Guardrails, enabling formal verification techniques to reduce AI hallucinations and ensure responsible GenAI outputs with up to 99% verification accuracy.

AWS has announced the general availability of Automated Reasoning checks in Amazon Bedrock Guardrails.

This new feature uses formal verification, a mathematically proven technique, to validate outputs from foundation models in real time. The feature enhances safety by minimizing hallucinations and incorrect responses, offering up to 99% verification accuracy.

First previewed at AWS re:Invent, this capability is now generally available and is part of AWS's broader push to provide secure, scalable, and responsible AI development through Bedrock.

Bedrock

AWS