IBM announced industry-first software that enables AI security and AI governance teams to collaborate on unified oversight of an enterprise’s risk posture.
This new capability strengthens and integrates watsonx.governance and Guardium AI Security to help clients operate AI systems, including agents, at scale while ensuring their security and reliability. watsonx.governance is IBM’s end-to-end AI governance tool, and Guardium AI Security is IBM’s tool for protecting AI models, data and usage.
“AI agents are poised to revolutionize enterprise productivity, but their benefits also pose challenges,” said Ritika Gunnar, general manager, Data & AI, IBM. “Without proper governance and security, these autonomous systems could do significant harm.”
Also Read: TOPPAN Unveils Its Generative AI-Powered Avatar Service
This new offering includes the following features:
Integrating and automating security for agent-based AI:
IBM is strengthening the integration between IBM Guardium AI Security and watsonx.governance to provide enterprises with the industry’s first integrated solution for managing the security and governance risks associated with AI use cases. The integration supports users in the process of validating compliance standards against 12 different frameworks, including EU AI regulations and ISO 42001.
IBM is also working with AllTrue.ai to introduce new capabilities to Guardium AI Security, including the ability to discover new AI use cases within cloud environments, code repositories and embedded systems, providing broader visibility and protection for an increasingly distributed AI ecosystem. Upon identifying a threat, IBM Guardium AI Security will automatically trigger the appropriate governance workflow via watsonx.governance.
The latest updates to IBM Guardium AI Security also include automated red teaming, allowing organizations to detect and remediate vulnerabilities and misconfigurations across AI use cases. Organizations can also define custom security policies that analyze both input and output prompts to mitigate risks such as code injection, sensitive data exposure and data leakage. These capabilities, currently available in IBM Guardium AI Security, and the integration with watsonx.governance are expected to roll out by the end of the year.
SOURCE: PRTimes
