Gergely Orosz says security certifications fall short against AI agent threats
Overview
Software engineer Gergely Orosz says security certifications are "less than useless" against threats from AI agents, in a post on X dated October 31, 2026.
He argues that certifications such as SOC 2 Type 2 and ISO 27001 do not test real-time monitoring, anomaly detection on latency or unauthorized access attempts, or continuous posture assessment by agents. As an example, he cites Binance, which holds both certifications. The post is a single claim by Orosz; the report does not include a response from Binance or from certification bodies.
Written by AI from the articles below · updated Oct 31, 11:10 AM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
Gergely Orosz@GergelyOroszXGergely Orosz says security certifications fall short against AI agent threatsAIGergely Orosz says security certifications are less than useless in the new AI agent threat landscape. He argues they do not check realtime monitoring, anomaly detection on latency or unauthorized access attempts, or continuous posture assessment by agents. He cites Binance, which holds SOC 2 Type 2 and ISO 27001 certifications, as a company that has them.
Heat Trend
Not enough continuous observations to show a trend yet.