Google Cloud has announced a suite of advanced security enhancements at the 2025 Security Summit, aimed at fortifying AI ecosystems and leveraging artificial intelligence to elevate organizational defenses.
These updates focus on proactive vulnerability detection, automated threat intelligence processing, and workload optimization for security teams.
Central to the announcements is the expansion of Security Command Center’s AI Protection solution, which now integrates Sensitive Data Protection for discovering and redacting sensitive inputs in AI agents to counter data poisoning attacks.
Complementing this, Model Armor extends its screening mechanisms to mitigate prompt injection and tool poisoning in agent-tool interactions, including those with Model Context Protocol (MCP) servers.
Building on these foundations, Google introduced three key capabilities for securing agentic AI within Google Agentspace and Google Agent Builder.
The expanded AI agent inventory and risk identification feature, set for preview later this year, enables automated discovery of AI agents and MCP servers, facilitating the identification of vulnerabilities, misconfigurations, and high-risk interactions across hybrid environments.
Additionally, advanced in-line protection and posture controls extend Model Armor’s real-time safeguards to Agentspace prompts and responses, defending against runtime threats like jailbreaking and sensitive data exfiltration.
Specialized posture controls, also slated for preview, enforce security policies and compliance standards tailored to AI agents.
Proactive threat detection enhancements in Security Command Center leverage Mandiant’s frontline intelligence to surface anomalous behaviors and external threats, enhancing incident response through integrated detections for agent assets.
Advancing Agentic SOC
Google’s vision for an agentic Security Operations Center (SOC) takes center stage with the introduction of AI-driven agents that coordinate actions to optimize data pipelines, automate alert triage, and streamline detection engineering.
The Alert Investigation agent, now in preview, autonomously enriches security events, analyzes command-line interfaces (CLIs), and constructs process trees based on Mandiant’s analyst best practices.
This agent delivers verdicts with actionable recommendations, significantly reducing manual toil and mean time to response (MTTR) in threat hunting and incident management.
Mandiant Consulting expands its offerings with risk-based AI governance, pre-deployment hardening guidance, and threat modeling to address customer needs for robust frameworks, adversarial testing, and detection mechanisms in generative and agentic AI deployments.
These services help organizations mitigate risks associated with AI adoption, including supply chain vulnerabilities and model exploitation.
In Google Unified Security, a converged platform powered by Gemini AI and Google’s global security fabric, new features include SecOps Labs for early access to AI-enhanced parsing, detection, and response experiments.
Generally available SOAR data integration dashboards provide visualization and analytics for security data, informed by customer feedback.
Chrome Enterprise enhancements extend secure browsing controls to iOS, enabling account separation, URL filtering to curb shadow AI risks, and detailed reporting for security events across mobile devices.
Innovations in Trusted Cloud Infrastructure
Updates to Trusted Cloud further strengthen compliance and access management. Security Command Center’s Compliance Manager, in preview, unifies policy enforcement, monitoring, and auditing with AI-specific controls for workload compliance.
Data Security Posture Management integrates with BigQuery for in-console monitoring, while Risk Reports use virtual red team simulations to highlight cloud security gaps.
Identity and Access Management (IAM) introduces Agentic IAM for auto-provisioning agent identities with observability, IAM role picker powered by Gemini for least-privilege recommendations, and re-authentication for sensitive actions.
Data Security expansions include Sensitive Data Protection for Vertex AI and image inspection, alongside Cloud Key Management System Autokey for streamlined encryption.
Network Security advancements feature org-scope tags in Cloud NGFW, support for RDMA networks in high-performance computing, and hierarchical policies in Cloud Armor Enterprise with enhanced WAF rules and rate limiting based on JA4 fingerprints and ASN intelligence.
These innovations underscore Google’s commitment to responsible AI deployment, automating compliance and empowering defenders against evolving threats in a landscape where AI redefines security postures.
Find this News Interesting! Follow us on Google News, LinkedIn, and X to Get Instant Updates!
Source link
