• Google Cloud unveils tools to secure AI agents at 2025 Security Summit.
• Updates help keep AI projects compliant.

Google Cloud used its Security Summit 2025 to highlight a wave of new tools aimed at protecting artificial intelligence projects while also using AI to strengthen cybersecurity operations.

The company pointed to the dual challenge facing organisations: securing their AI systems and applying AI to defend against threats. The announcements covered updates in Google’s security portfolio, including Security Command Center, Google Security Operations, Chrome Enterprise, and Trusted Cloud.

Safeguarding AI agents

As AI agents become more common in enterprise environments, Google Cloud introduced new capabilities in its Security Command Centre to monitor and protect them. A preview feature will give teams automated discovery of AI agents and Model Context Protocol servers, flagging vulnerabilities and risky interactions.

Google is also expanding its Model Armor tool to cover prompts and responses in Agentspace, offering real-time defences against threats like prompt injection or jailbreaking. New posture controls for Agentspace and Agent Builder are also on the way, designed to enforce organisational security policies.

To help with ongoing protection, new detections have been added to identify suspicious behaviour in AI agents, drawing on intelligence from both Mandiant and Google.

AI in security operations

The company also shared its vision for an “agentic” security operations centre, where AI-driven agents can coordinate tasks to support human analysts. An Alert Investigation agent, now in preview, can automatically enrich events, analyse command-line inputs, and build process trees, with recommendations for next steps. Google said this could ease manual workloads and reduce response times.

Mandiant Consulting, part of Google Cloud, is expanding its services to address security concerns raised by the rapid use of generative and agentic AI. The firm is offering support in AI governance, environment hardening, and threat modelling. It has also emphasised the role of adversarial testing to identify risks before deployment.

Updates in Google Unified Security

The company’s broader platform, Google Unified Security, also received several upgrades. A new SecOps Labs feature in Google Security Operations will give customers early access to AI-powered experiments for detection and response. New dashboards, now generally available, integrate SOAR data for improved visibility.

Chrome Enterprise added new mobile protections, including URL filtering on iOS and enhanced reporting to track browser activity. The features are intended to help organisations manage employee browsing risks, including use of unapproved generative AI tools.

Cloud security enhancements

On the cloud side, Google announced previews of Compliance Manager and Data Security Posture Management in Security Command Center. The tools aim to simplify compliance monitoring and sensitive data governance, including AI-specific workloads. Risk Reports, also in preview, will summarise potential weaknesses based on simulated attacker techniques.

Identity and Access Management is adding “agentic IAM” later this year, which will provision agent identities in environments. Other updates include a role picker that uses Gemini to suggest least-privilege roles, and re-authentication requirements for sensitive actions.

Data security updates include expanded scanning of AI-related assets and image data, while network security improvements extend Cloud Armour and Cloud NGFW features for broader protection.

Google security focus on responsible AI adoption

The company framed these updates as part of its effort to make AI adoption more secure while using AI itself to enhance defence. By embedding controls into data protection, compliance, and identity systems, Google Cloud said it aims to give organisations a stronger foundation for managing both risks and opportunities in AI.

Want to learn more about cybersecurity and the cloud from industry leaders? Check out Cyber Security & Cloud Expo taking place in Amsterdam, California, and London.

Explore other upcoming enterprise technology events and webinars powered by TechForge here.