SEATTLE — Amazon Web Services (AWS) announced the launch of "AWS Continuum" on Tuesday, a new automated security tool designed to identify and remediate software vulnerabilities. The release comes amidst growing concerns that advanced artificial intelligence models, such as Anthropic’s "Mythos," possess the capability to detect and exploit cybersecurity flaws with expert-level precision.

AWS Continuum acts as an end-to-end security engine. It performs deep code analysis to identify potential vulnerabilities, categorizes them by risk level, and automatically generates patches to address them. A key feature of the tool is its ability to verify vulnerabilities within an isolated environment by simulating real-world attack scenarios, ensuring that detected threats are genuinely exploitable before taking action.

The service offers two distinct operational modes: a "Learning Mode," which allows for human oversight and verification, and an "Enforcement Mode," which handles remediation autonomously. AWS recommends that organizations begin with Learning Mode to build confidence in the system's recommendations before transitioning to full automation.

Furthermore, AWS has integrated the penetration testing and code inspection capabilities of its existing security solution, "AWS Security Agent," into the new platform, creating a more robust defense ecosystem.

Industry experts note that the launch is a strategic response to the rapid evolution of high-performance AI models like Anthropic’s "Claude" and OpenAI’s "GPT," which have accelerated the speed at which vulnerabilities can be discovered—and potentially weaponized.

Amazon had previously alerted the U.S. government regarding concerns that models such as "Claude Fable 5" and "Claude Mythos 5" could be manipulated for cyberattacks. This warning is widely believed to have played a significant role in the recent temporary suspension of those specific models.

As cyber threats become increasingly AI-driven, AWS Continuum represents a proactive step in equipping developers and security teams with the tools necessary to stay ahead of automated adversaries.

[ⓒ Global Economic Times. 무단전재-재배포 금지]