Hack The Box says 68% of top 25 cyber teams used AI agents

Hack The Box released a 2026 benchmark showing faster solve times, more full-board completions and broader AI agent use among elite cybersecurity teams. The report says 17 of the top 25 teams had AI agents, even as human judgment remained central to results.

Why it matters: - AI is moving into day-to-day cybersecurity workflows, especially among top-performing teams. - The report suggests stronger teams are pairing AI tools with human judgment, validation and hands-on technical skill. - Faster solve times and more full-board completions point to rising team capability across the competition.

What happened: - Hack The Box released its 2026 Global Cyber Skills Benchmark Research Brief on Aug. 26, 2026. - The three-year analysis tracked performance across the competition and compared results from 2024 through 2026. - AI agent accounts made up 2.7% of all registered accounts. - Seventeen of the top 25 teams, or 68%, had an AI agent. - Those agents accounted for 4.2% of submitted flags and 4.6% of points awarded. - The findings do not prove AI caused better performance.

The details: - Median recorded time-to-solve fell from 26.1 hours in 2024 to 13.8 hours in 2026. - That drop was more than 12 hours over the three-year period. - Teams completing the full challenge board rose from two in 2024 to three in 2025 and 15 in 2026. - The board expanded even as completions increased. - The report links the findings to a broader shift in cybersecurity, where AI is affecting both attack and defense. - Hack The Box cited recent industry disclosures, including Hugging Face’s July 2026 incident disclosure and OWASP’s Q1 2026 GenAI exploit roundup, as examples of that shift. - The company said the core challenge for security leaders is not just adopting AI, but ensuring practitioners can direct, test and validate its work. - Haris Pylarinos, founder and CEO of Hack The Box, said AI is already part of cybersecurity operations on both sides of the equation.

Between the lines: - The data points to AI becoming a practical tool for elite practitioners, not a replacement for them. - The concentration of AI agent use among top teams suggests the strongest operators are more likely to experiment with, and integrate, automation. - The results also reinforce a skills gap concern: faster AI systems can raise the value of human oversight rather than reduce it. - Hack The Box is framing the benchmark as evidence that cybersecurity talent development now has to include agent use, validation and decision-making.

What’s next: - Hack The Box says the full 2026 Global Cyber Skills Benchmark Research Brief is available here. - The company is positioning its platform around training and validating both humans and AI agents. - Future benchmarks will likely show whether AI adoption continues to spread beyond top teams and whether performance gains keep accelerating.

The bottom line: - AI agents are already part of the toolkit for many of the best cybersecurity teams, but the report says human expertise still drives how effectively those tools are used.