Anthropic’s Mythos AI Sparks Cybersecurity Panic, but Experts Say the Threat Already Exists

Anthropic’s new AI model, Mythos, triggered panic across the cybersecurity industry after reports claimed it uncovered thousands of unknown software vulnerabilities. But cybersecurity researchers say the real story is far less sudden because the same kind of attacks were already possible using older AI models from both Anthropic and OpenAI.



The concern around Mythos centers on how quickly AI can now discover “zero-day” vulnerabilities, which are hidden software flaws that attackers exploit before developers release patches. Researchers say the industry already entered this phase months ago, but Mythos forced companies and governments to finally confront it.



Ben Harris, CEO of cybersecurity firm watchTowr, described the reaction from banks and regulators as “hysteria” after Mythos arrived. He explained that researchers are already reproducing many of Mythos’ results using public AI models and orchestration techniques that coordinate multiple tools together.




“The industry is panicking about the number of vulnerabilities they face now. But even before Mythos is widely available, it couldn’t fix vulnerabilities fast enough.”— Ben Harris, CEO of watchTowr




That gap between discovering vulnerabilities and patching them has become the industry’s biggest problem. AI now finds security flaws in hours, while companies still need days or weeks to release fixes, especially when critical systems require downtime before updates can happen.



Klaudia Kloc, CEO of Vidoc, said researchers successfully recreated Mythos-style vulnerability discoveries using older Anthropic and OpenAI models. She warned that current AI systems are already “scary enough” because they can detect zero-days at scale.



Another major concern involves automation. Anthropic says Mythos can move beyond finding vulnerabilities and begin creating working exploits with very little human input. Cybersecurity experts argue that skilled hackers from countries like North Korea, China, and Russia already possess those capabilities with existing tools.



While companies race to build AI-powered defenses, researchers say attackers currently hold the advantage because offensive AI tools are evolving faster than defensive systems. That imbalance explains why Mythos sparked so much fear across the cybersecurity world, even though experts insist the threat was already here.