ServiceNow Issues Patches for Four Security Flaws in AI Platform

ServiceNow announced today it had deployed a security update addressing four distinct flaws in its AI Platform. Three received a maximum CVSS rating of 10.0, indicating they can be exploited by attackers without authentication to run arbitrary code or inject malicious SQL commands against the underlying PostgreSQL database engine.¹ The company confirmed that updates were pushed to all hosted instances and distributed to partners and self-hosted customers, noting that older deployments remain exposed until patched.

The affected component is ServiceNow’s JavaScript‑based orchestration engine, which executes scripts triggered by events within the platform. According to release notes, bugs stem from insufficient sanitization of input to this scripting environment, allowing elevated‑privilege code execution if successful in either hosted or on‑premises installations where updates are not applied promptly after release. Similar vulnerabilities—such as CVE‑2023‑46897—were previously found in the same product line and led to widespread exploitation targeting enterprise customers before patches became available for systems lacking timely update cycles.

Experts note the high severity reflects both ease of exploitability and potential impact on data integrity across many organizations dependent upon their use cases involving sensitive information stored using ServiceNow’s cloud services.

In practical terms, an attacker could gain root access to a customer’s entire database instance via crafted inputs, allowing them to read, modify, or delete records. The AI Platform is widely used by enterprises for incident response automation, change management, and other critical workflows; any compromise could lead to significant privacy breaches. Immediate patching is advised, especially since the vulnerability does not require valid credentials—just network connectivity—to be exploited.

ServiceNow also highlighted that it has increased monitoring and testing procedures following these findings, urging customers to verify all updates have been applied and review audit logs for anomalous activity. While patches are now available, large deployments may experience delayed rollout due to complex testing cycles before updates occur.

The disclosure underscores how well‑known platforms can harbor severe flaws threatening operational continuity and data confidentiality when left unpatched. It serves as reminder for organizations relying on third‑party cloud services to maintain rigorous update schedules and conduct regular penetration tests focused on input sanitization paths within their orchestration engines.