Print blog article
Blog
Is the pentester the next dodo of the tech industry?
With last week's launch of Claude Code Security, Anthropic makes one thing clear: AI doesn't stop at writing code; it's also starting to cannibalize the role of cybersecurity auditor.
24 February 2026 | 1 minute read
We saw it coming. The question is how fast it will develop.
The rise of “agentic security”
When an AI agent identifies vulnerabilities and applies patches in real-time as code is being written, it changes the entire industry.
From “Shift-Left” to “Always On”
Cybersecurity is no longer a check at the end of the ride, but an integral part of your pipeline.
“The death of low-hanging fruit”
The standard OWASP Top 10 findings that used to fill reports? AI now catches them before the code is even committed.
Immediate defense
Real-time feedback on complex risks, such as prompt injection or unsafe data handling during the development process.
The new reality?
The “commodity” pentester who relies on scanners and known patterns has a problem. The human expert is forced to evolve into what AI cannot (yet) do: cracking complex business logic and devising creative attack chains that fall outside the predictable framework of an LLM.
The bar for secure software may just have been raised significantly. The question is not whether penetration testing will disappear, but whether pentesters will evolve quickly enough.
Or is it all nonsense? I don't think so.
This post was first published on LinkedIn and translated from Dutch.
