- Google Cloud launches New AI Protection Security Suite
- Offering identifies, assesses and protects AI assets for vulnerabilities
- Even more security features will come soon
Google Cloud has launched AI Protection, a package of security features designed to mitigate risks across AI work loads and data, regardless of the platform used.
The new offering will give companies a centralized view of their AI -standing so that they can control the risks and mockery of the spotting threats before becoming a problem.
“When AI uses rise, security remains a top concern and we often hear that organizations are concerned about risks that can come up with quick adoption,” noted Archana Ramamoorthy, senior director, product management, Google Cloud Security. “Google Cloud is required to help our customers with confidence build and implement AI in a safe, compatible and private way.”
Increased protection of AI workloads
AI protection will be built into the Security Command Center (SCC), which provides a centralized AI security management system along with other cloud -risks.
Among the new platform’s central capacity is AI-Lager Discovery (identifies and evaluates AI assets for vulnerabilities), AI active safety (implements control, policies and Roger to secure AI resources) and threat management (offering detection, investigation and response mechanisms for AI-related threats).
Furthermore, Google Cloud explained that its sensitive data protection (SDP) improvements are now expanded to Vertex AI data sets, enabling automatic discovery and classification of sensitive training and setting data. After discovering sensitive data, AI protection will use SCCS virtual red teaming to identify potential attacking paths on AI systems and suggest remedy steps.
Google Cloud also said that Model Armor, a core capacity for AI protection, is now generally available. It is designed to protect against fast injection and jailbreak attacks, data loss and malicious URLs and offensive content. It can be integrated into applications via residual API, Apigee and soon Vertex AI.
Finally, AI Protection will operationalize security information and research from both Google and Mandiant to help defend AI systems.
Initial access trials, privilege -scaling and persistence experiments on AI workloads can all be detected via SCC, while new detectors for AI protection, based on the latest front line -Intelligence, “Coming Soon”. These will help identify and manage Runtime threats such as basic model hijacking.
