Amazon Bedrock has launched an updated ApplyGuardrail API code example, enhancing safeguards for generative AI applications. The Guardrails now offer customizable protections in addition to native safeguards, setting a new industry standard for safety features. The update introduces contextual grounding checks to detect hallucinations in model responses and the ApplyGuardrail API to evaluate input prompts and model responses for all foundation models. MAPFRE, Spain’s largest insurance company, has already implemented Guardrails to align with its security policies and responsible AI practices. The new capabilities are available in all AWS Regions where Guardrails for Amazon Bedrock is available. This update reflects Amazon Bedrock’s commitment to providing advanced safeguards and governance for generative AI applications, empowering businesses to develop innovative, secure, and responsible AI practices. For more information, visit the Guardrails for Amazon Bedrock product page, the Amazon Bedrock pricing page, or the community.aws site for technical content and solutions.
The 128 series brings you byte-sized news for busy professionals. The full story can be found here: https://cmd.news/article/industry-news/amazon-bedrocks-guardrails-update-applyguardrail-api-and-contextual-grounding-check/
