Anthropic Mythos: AI Vulnerability Threat Is Already Here

A new AI model called Mythos from Anthropic has caused alarm among global banks, tech giants, and governments. Here's the thing: Mythos reportedly found thousands of unknown software vulnerabilities. What's interesting is, cybersecurity experts say this capability isn't new. Existing AI models, including those from Anthropic and OpenAI, can already uncover these same weaknesses. For example, Ben Harris, CEO of watchTowr Labs, says people are reproducing Mythos's findings with public models. Anthropic limited Mythos’s release to companies like Apple and JPMorgan Chase, but the Trump administration is now considering new AI oversight. The bottom line: This isn't a future threat; the ability for AI to find software flaws at scale is already here, meaning everyone needs to strengthen their digital defenses now.