Alex Stamos, chief security officer at Cognition, told The Verge that the multibillion-dollar cybersecurity sector is being overwhelmed by AI-enhanced hacking campaigns, describing the situation as a looming "dark age" for digital defenses. He said AI is now enabling both private criminals and nation-state actors to locate and exploit flaws at unprecedented speed.
In the past month, a series of linked intrusions have illustrated the new threat pattern. Vulnerabilities in widely used platforms such as PHP and Node were exploited, followed by the high-profile Binance theft, which Stamos called the largest financial heist in history. A breach of the AI service provider Fireworks exposed API keys that were used to compromise the Cursor platform, which then distributed a malicious plugin to roughly 5,000 downstream firms.
Stamos explained that adversaries have reached a point where open-weight models can automate the entire cyber "kill chain",the sequence of steps from reconnaissance to data exfiltration. He said these systems can scan targets, generate exploits, and launch attacks in real time, eliminating the traditional need for human-crafted code before deployment.
He compared the current turmoil to the late-1990s Windows security crisis, noting that back then the internet was not yet ubiquitous. "Every company in the world felt vulnerable," Stamos recalled, adding that the scale of impact today is far greater because the entire global economy now runs online and can be compromised automatically.
The proliferation of openly available AI models without built-in logging or safeguards has lowered the barrier to entry for attackers. Stamos highlighted that tools once limited to a small community of experts can now be run on consumer-grade hardware, allowing malicious actors in jurisdictions beyond law-enforcement reach to launch sophisticated campaigns.
Historically, the industry invested heavily in bug-bounty programs and paid security researchers to protect systems. Stamos warned that this economic model is eroding as inexpensive AI tools enable attackers to generate exploits cheaply, shifting the cost balance in favor of offense rather than defense.