Former NSA hacker David Kennedy said a swarm of AI agents breached a company’s systems in 14 minutes. The attackers used valid credentials to bypass two-factor authentication and steal sensitive data. "Never seen anything that fast," Kennedy said. The company did not detect the breach for over four hours.
Hackers are now using swarms of AI agents to attack companies and steal data. Former US NSA hacker David Kennedy claims that in a cyberattack he is defending, AI managed to breach the systems in a matter of minutes.
Former US National Security Agency (NSA) hacker David Kennedy has revealed that a company was recently targeted by a swarm of AI agents. These AI agents managed to breach the company's systems and steal sensitive data in a matter of 14 minutes, something Kennedy had never seen before. "It all took 14 minutes," David Kennedy told CNN. "Never seen anything that fast."
David Kennedy, who now works in cybersecurity, says that one of his clients recently contacted him about a cyberattack. He said that the affected company did not even notice that it had been compromised for over four hours. "They didn't even detect it for about four and a half hours until the system was then knocked offline," he explained.
Though he did not name the organisation that was breached.
What was the AI swarm attack?
Explaining how the attack unfolded, David Kennedy said the way hackers work has changed, with different ransomware groups having different levels of expertise. Some groups tried to "phish and harvest credentials, usernames and passwords from companies" before selling them to more sophisticated attackers capable of targeting their systems.
In this case, the hackers used valid usernames and passwords before launching what Kennedy called an "agent swarm." A swarm of AI agents refers to a group of different AI agents that usually target different components of a company's digital infrastructure. "We saw over 30 different types of agents probing their infrastructure at the exact same time," he added.
AI agents bypass two-factor authentication
One of the stricter safety measures is two-factor authentication that ensures that an unauthorised person, or agent, does not get access to systems. But Kennedy said that the AI agent swarm had managed to bypass this through a flaw in the company's systems. The agents were "able to find a flaw in their application on how they use what's called two-factor authentication. It bypassed that. They got access to the system," the former NSA hacker explained.
After gaining access, the AI agents were able to steal a lot of data. The hackers then deployed ransomware that encrypted the company's entire file system. "And this all took 14 minutes. 14 minutes," Kennedy said. The system initially continued operating before encountering errors and restarting, which led the company to notice an outage hours later.
This incident comes at a time when cases of AI-driven cyberattacks are growing. Google's Threat Intelligence Group recently warned that hackers had developed an AI framework capable of scanning security patches, identifying flaws in older software versions, and planning and carrying out attacks independently. Microsoft too has identified China-linked hacker groups that are using AI tools for cyberattacks.
Previously, there have been cases of AI agents going rogue and hacking companies without authorisation. OpenAI's infamous Hugging Face breach saw over 700 AI agents try to hack the US company.
