Hack The Box (HTB) has published findings from its latest AI-Augmented vs. Human-Only Cybersecurity Performance Benchmark Report. The research draws on data from its NeuroGrid Capture The Flag (CTF) competition, which compares AI-augmented and human performance on cybersecurity tasks.
The results show that AI integration can increase task completion speed depending on the proficiency of the AI-augmented team. Key findings indicate that AI-enabled teams completed tasks faster, generating up to 4.1x more output for elite teams and 1.4x more across all teams within a set timeframe.
AI-augmented teams also recorded a 70% higher challenge solve rate compared to top human-only teams, achieving a 3.2x higher solve-rate ratio across all participants.
The benchmark included 1,078 teams — 120 agentic AI teams and 958 human teams — participating in 36 cybersecurity challenges across nine technical domains and four difficulty levels over a three-day period.
The report highlights that the effects of AI adoption vary across experience levels and that workforce development strategies may need to account for these differences:
- Early Career: AI can support less experienced teams in completing more challenges but may also lead to reduced efficiency in some cases. These teams were observed to be 12.5% slower on average, sometimes becoming dependent on iterative or unstructured workflows without sufficient oversight.
- Mid Career: Mid-level operations saw the strongest improvement on medium-difficulty tasks, with a peak advantage of 3.89x.
- Elite Teams: While the relative advantage in solve rate narrows at higher experience levels, AI-augmented elite teams demonstrated a speed increase, completing challenges 312% faster.
The findings suggest that automation of routine and mid-level tasks can deliver measurable productivity gains. At the same time, the report notes that over-reliance on automation in judgement-based tasks may affect long-term skill development and resilience. The competitive advantage lies not only in adopting AI tools, but also in developing the capability to effectively manage, validate, and govern AI-driven workflows within cybersecurity operations.
