Champion ethical hacker warns AI tools like Mythos will make competing harder
Champion Ethical Hacker Warnings: AI Tools Like Mythos Could Reshape Cybersecurity Competition
Champion ethical hacker warns AI tools – The cybersecurity landscape is undergoing a significant transformation due to advancements in artificial intelligence, according to Valentina Palmiotti, a celebrated ethical hacker who recently secured top honors at the annual Pwn2Own hacking competition in Berlin. Known by her online alias Chompie, she highlighted concerns that AI-driven tools such as Claude Mythos might soon render human competitors obsolete, challenging the status quo in the field of vulnerability discovery.
A New Era in Hacking Competitions
Pwn2Own, organized by the ZeroDay Initiative, has long been a proving ground for ethical hackers worldwide. This year, the event awarded nearly $1.3 million in prize money to participants who uncovered 47 previously unknown security flaws across software, websites, and systems. These vulnerabilities, now patched by the affected companies, were critical to preventing cybercriminals from exploiting them. Yet, as the competition progresses, the role of AI is becoming increasingly prominent, raising questions about its impact on human hackers.
Chompie, who emerged as the most successful individual at the 2024 Pwn2Own, described how AI tools are already altering the game. She noted that while these technologies are currently aiding her in “bug bounties”—cash rewards for identifying weaknesses in digital systems—future iterations like Mythos may soon outpace even the most skilled human hackers. “It’s not just about speed anymore,” she remarked, “but about how much AI can replace the human element in problem-solving.”
The Human Edge in a Digital Age
At the heart of the competition lies a blend of human ingenuity and machine precision. Chompie’s victory on the first day of the event, where she successfully breached an Nvidia-linked system to claim $20,000, was followed by what she called “zombie hacker mode.” This state of intense focus, sustained by energy drinks and adrenaline, saw her working through the night to prepare for the next challenge. “I didn’t sleep at all,” she admitted, “but it was worth it.”
Footage of her final demonstration, which earned her $50,000, captured the culmination of her efforts. She hacked into a Linux-based system, showcasing her expertise in a high-stakes environment. However, her achievements were not solely the result of her own skills. She revealed that tools like Claude Code had become essential in accelerating her work during competitions and in her professional capacity as a security researcher for IBM X-Force. “AI is like having a supercharged assistant,” she said, “but it’s only getting better.”
The Threat of Mythos and Emerging Models
Mythos, developed by Anthropic, has sparked particular attention for its ability to uncover 1,600 vulnerabilities across hundreds of software programs. This capability has led to speculation about its potential as a tool for both defenders and attackers. Anthropic claims the model’s power is so great that it is restricted to a select group of governments and cybersecurity institutions, underscoring its strategic importance.
Chompie believes that as AI models evolve, the competition for top-tier hacking prizes will intensify. She predicted that new systems like Mythos and GPT 5.5 Cyber will soon make it difficult for even the best human hackers to keep up. “I competed this year because I thought it might be my last chance,” she explained. “It’s not that I’m pessimistic about the future of ethical hacking, but I see the lower-hanging fruit disappearing.”
Human Creativity vs. Machine Efficiency
While Chompie expresses apprehension about the rise of AI, not all hackers share her perspective. Orange Tsai, a Taiwanese security researcher and one of the event’s top winners, views AI as a valuable ally rather than a threat. His team claimed $375,000 by discovering intricate attack vectors, a feat he attributes to the synergy between human intuition and machine learning. “AI feels more like an awesome assistant,” he said, “helping me accelerate my workflow. I still need to sleep, but it can handle the repetitive tasks.”
Orange Tsai emphasized that while AI raises the bar for security research, it also opens new avenues for human innovation. “There are still vulnerabilities that only a human mind can spot,” he argued. “Creativity and intuition will always be key, even if AI helps us test more efficiently.” His optimism contrasts with Chompie’s cautious outlook, highlighting a divide in how the community is responding to the integration of AI into hacking.
The Broader Implications for Cybersecurity
The growing influence of AI extends beyond ethical hacking, with criminal actors also adopting these tools to enhance their operations. Research indicates that cybercriminals are using AI to automate attacks, identify new attack paths, and execute data breaches with unprecedented speed. This trend could reduce the time required to exploit weaknesses, putting more pressure on defenders to stay ahead.
Yet, most cyberattacks still rely on established methods like phishing and social engineering. These techniques, which exploit human behavior rather than technical complexity, remain a staple for attackers. Chompie noted that while AI may replace some of the easier vulnerabilities, the harder ones—those requiring deep analysis and innovative thinking—will still require human expertise. “The real challenge is in finding the bugs that AI hasn’t seen yet,” she said, “and that’s where the top hackers will still shine.”
A Future of Specialization and Adaptation
As AI tools like Mythos mature, Chompie suggested that the cybersecurity field may shift toward specialization. She envisioned a future where only the most elite hackers, capable of navigating AI-generated complexity, would remain in the spotlight. “People like Orange Tsai are already pushing the boundaries,” she observed, “and they’ll continue to do so even as AI becomes more dominant.”
Despite the challenges, Chompie remains confident in the resilience of ethical hacking. She believes that while AI will streamline processes, it won’t eliminate the need for human insight. “The competition is getting tougher, but that’s also what makes it exciting,” she concluded. “We’re evolving, and the best hackers will adapt to stay relevant.”
Orange Tsai echoed this sentiment, stressing that AI’s role is to complement, not replace, human ingenuity. “It’s a tool, not a replacement,” he said. “We’ll still need people to think outside the box and imagine what AI hasn’t already tried.” For now, the balance between human and machine remains precarious, but both Chompie and Tsai agree that the future of cybersecurity will be defined by those who can harness the power of AI without losing their edge.
As the competition continues to evolve, the question remains: will AI redefine the skills required to succeed in cybersecurity, or will it simply elevate the standards set by human pioneers? For Chompie, the answer is clear. “It’s a new chapter,” she said. “But the core of what we do—finding weaknesses and protecting systems—will always matter.”