Overview
- Google disclosed on May 27 that it has launched AI Threat Defense, a four-stage enterprise product that prepares environments, scans and prioritizes findings, remediates vulnerabilities, and monitors runtime systems.
- The service stitches together Gemini reasoning, Wiz cloud-exposure mapping, DeepMind’s CodeMender for automated fixes, and Mandiant threat intelligence and incident response to run a single, automated workflow.
- CodeMender creates patches inside developers’ IDEs or CLIs, can rewrite legacy code into memory-safe languages, generates tests to verify fixes, and tags patched libraries in source control before deployment.
- Prioritization uses Wiz to check real-world reachability so isolated high-severity flaws are deprioritized, and monitoring relies on Google Security Operations agents plus daily-built signed container images to limit runtime risk.
- Google announced partner channels with TENEX.AI and Netenrich and positioned the product against Anthropic’s Mythos and OpenAI’s Daybreak, while analysts warn that autonomous remediation raises operational and governance risks that require human oversight.