How Standard Bank Group automates AI risk assessment for generative AI
Standard Bank Group automated the evidence behind its generative AI approval gates. Working with AWS Professional Services, the bank built an assessment in its own AWS account that scores each use case against international AI standards and its own policies using LLM-as-a-judge on Amazon Bedrock, giving risk committees consistent, traceable evidence.
Standard Bank Group has automated AI risk assessment for generative AI applications using AWS Professional Services. This automation addresses the slow approval process caused by manual reviews and proof gathering. Standard Bank already had a model risk practice and change risk process, which assess each new use case's risk before it becomes live.
The new assessment automates evidence gathering and testing, ensuring consistent evidence collection across use cases, which streamlines the approval gates. The assessment utilizes an LLM-as-a-judge technique with Amazon Bedrock, which evaluates the evidence against defined criteria and controls covering international AI standards and the bank's own policies.
The automated assessment consists of four steps: evidence collection, artifact examination, application testing, and evaluation summarization. The assessment runs entirely within the bank's AWS account using Amazon S3, AWS Well-Architected Tool, AWS Step Functions, and Amazon Bedrock. The bank retains control over the foundation model through AWS IAM, preventing unauthorized access to model prompts and outputs.
Written by urgent.news from AWS Machine Learning's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
This story
This is one outlet's version. Read the fullest account.