The rule 'Unusual High Word Policy Blocks Detected' is designed to monitor and detect repeated compliance violations related to the Amazon Bedrock generative AI service specifically focusing on a policy termed 'word_policy'. This policy is activated when content flagged as inappropriate or against the guidelines is attempted. The detection works by querying logs from AWS Bedrock for instances where the policy action is marked as 'BLOCKED' alongside a recorded violation. The rule counts how many times an individual user has triggered this policy and triggers an alert when a user exceeds five violations within a defined time frame. This could indicate either misconduct or an account that is being probed for its denial of access to certain content topics. The rule employs a 10-minute interval check over the past hour to ensure timely responses to potential threats. In cases of false positives, such as new model deployments or compliance testing that might trigger legitimate blocks, precautions are suggested to validate the context before taking action.