AI Is Speeding Up Cloud Attacks: What CrowdStrike's New Cloud Security Tools Aim to Address

AI is accelerating both sides of cloud security. According to CrowdStrike, developers and coding agents are writing more code, pulling in more open source packages and launching more cloud and AI workloads than security teams can review by hand. Attackers are also using AI to move faster, and they are targeting software packages in automated pipelines, the identities that control cloud resources, and the infrastructure behind AI applications.

CrowdStrike says the latest Falcon Cloud Security features are now generally available. They are designed to help organisations stop compromised software before it reaches production, protect AI workloads and supporting infrastructure, and detect cloud attacks in seconds. The company says the product is backed by intelligence on more than 290 adversaries and shares one platform with its endpoint and identity protection, so the risk context built before an attack can be used to respond during one.

The software supply chain is a particular concern. CrowdStrike reports that it observed the adversary ALTERED SPIDER compromise more than 300 software dependencies in a single day. Because AI coding agents can pull open source packages into automated builds faster than people can check them, attackers can exploit that trust by compromising otherwise legitimate software. CrowdStrike notes that a compromised dependency does not need a published vulnerability (CVE) to be dangerous, so vulnerability scanning alone can miss the threat.