The Red Hat security advisory RHSA-2025:16982 announces the release of Red Hat OpenShift AI 2.22.2, which addresses two security vulnerabilities identified as CVE-2025-10725 and CVE-2025-57852. These vulnerabilities involve issues categorized under CWE-266 (Incorrect Privilege Assignment) and CWE-276 (Incorrect Default Permissions), indicating potential weaknesses in access control or authorization mechanisms. The advisory does not provide detailed technical descriptions or CVSS scores but classifies the severity as high. Updated container images for multiple architectures have been published to remediate these issues. No known exploits in the wild have been reported. The vendor provides documentation for upgrading clusters to apply the fixes fully.

The vulnerabilities could allow unauthorized access or privilege escalation due to incorrect access control or permissions within Red Hat OpenShift AI. This may lead to unauthorized actions within the affected environment. However, no active exploitation has been reported. The impact is considered high based on the vendor's severity rating.

Red Hat has released updated container images and the 2.22.2 version of Red Hat OpenShift AI to address these vulnerabilities. Users should follow the official Red Hat documentation to upgrade their clusters and apply the errata update fully. No other specific mitigations or workarounds are indicated. Patch status is confirmed by the vendor advisory, and the fix is available in version 2.22.2.