This AI Hacking System Crushed 99% Of Elite Human Hackers

AI Hackers Are Winning

In a stunning display of artificial intelligence capabilities, Israeli startup Tenzai has developed an AI hacking system that has just crushed 99% of human competitors in six elite capture the flag (CTF) cybersecurity competitions. The AI, which uses customized versions of OpenAI and Anthropic models, went head-to-head against 125,000 of the world's top cybersecurity experts and came out on top.

The competitions, which regularly update with new sets of tricky challenges, are considered among the most prestigious testing grounds for cybersecurity talent worldwide. Tenzai's performance represents what experts are calling a "singularity moment" for AI hacking and marks a potential turning point in the cybersecurity industry.

According to a report by Forbes, the AI system demonstrated unprecedented capabilities in identifying and chaining together software vulnerabilities. This development has sent shockwaves through the cybersecurity community, raising questions about the future of human hackers in an AI-dominated landscape. The results were published in March 2026 and have since been widely discussed in tech circles.

How The AI Hacking System Works

Tenzai's cofounder and CEO Pavel Gurvich explained that the AI was surprisingly adept at combining exploits for software vulnerabilities - something that had previously been difficult to automate. The system leverages tailored versions of leading large language models from OpenAI and Anthropic to analyze, strategize, and execute complex hacking challenges in real-time.

"What we're seeing is the convergence of advanced AI reasoning with cybersecurity expertise," Gurvich told Forbes. "The AI doesn't just find vulnerabilities - it chains them together in ways that even experienced human hackers struggle to anticipate. This represents a fundamental shift in how we think about automated hacking capabilities."

The system was tested across six different CTF competitions, each presenting unique challenges that required creative problem-solving and rapid adaptation. Despite facing off against some of the best human cybersecurity talent from around the world, the AI consistently outperformed nearly all of its human competitors.

What This Means For Cybersecurity

Gadi Evron, founder and CEO of AI security company Knostic, says hackers have already had their "singularity moment." According to Evron, as reported by Forbes, it used to take days or weeks to go from discovering a software vulnerability to exploiting it. Now, AI can do this in minutes or even seconds. This marks a significant acceleration in the speed of cyberattacks and defensive responses alike.

This development has massive implications for both offense and defense in cybersecurity. On one hand, organizations can use AI to identify and patch vulnerabilities before malicious actors can exploit them. On the other hand, the same technology could be used by threat actors to launch more sophisticated attacks at an unprecedented scale.

The dual-use nature of AI hacking technology means that both legitimate security teams and cybercriminals now have access to powerful automation tools. This arms race is just beginning, and organizations must adapt their security strategies accordingly to stay ahead of evolving threats in an increasingly dangerous digital landscape.

The Future Of AI In Hacking

The competition results suggest that AI is rapidly approaching - or has already surpassed - human-level capability in certain cybersecurity domains. As AI models continue to improve, we can expect to see more AI-powered tools being used in both legitimate security operations and, unfortunately, in cyberattacks conducted by malicious actors.

For now, Tenzai's achievement serves as a wake-up call for the cybersecurity industry. Organizations need to start thinking about how to defend against AI-powered threats, and security professionals need to upskill to stay relevant in an increasingly automated landscape where traditional methods may no longer be sufficient.

The rise of AI hacking systems also raises important ethical questions about responsible disclosure and the democratization of hacking tools. As this technology becomes more accessible, the line between white-hat and black-hat hacking may become increasingly blurred. Security researchers and policymakers will need to work together to establish clear guidelines for the responsible development and deployment of such powerful AI systems in the coming years.