Research from Tenable highlights that 72% of AI cloud workloads contain critical vulnerabilities compared to 59% of non-AI workloads. The complexity of AI workloads, especially those running on Unix systems with multiple libraries, increases security risks, including data manipulation and leakage. Misconfigurations, termed 'jenga-style,' also contribute to security threats. For instance, many organizations have default overprivileged service accounts resulting in vulnerabilities. With the growing adoption of AI tools, attention to these security issues should be a top priority for IT leaders.
Almost 72% of AI cloud workloads have a critical vulnerability, highlighting significant security risks that need immediate attention from IT leaders as AI adoption increases.
The report indicates that the combination of AI workloads and Unix-based systems creates vulnerabilities due to numerous libraries and risky configurations, making data manipulation a critical concern.
Tenable's findings reveal that 91% of companies using Amazon SageMaker have risky default administrator privileges, exposing system-critical files to user alterations, which can lead to severe security breaches.
The 'jenga-style' cloud misconfigurations present a major issue as organizations layer AI services, often unaware of how these configurations can jeopardize their overall security.
Collection
[
|
...
]