Truth and Consequences: AI Bull(s) in Our OT China Shops

Truth and Consequences: AI Bull(s) in Our OT China Shops

Title history
  • Truth and Consequences: AI Bull(s) in Our OT China Shops — observed on 20261110-2022 UTC (YouTube)
🎙 Joshua Corman 👥 71K 📅 November 10, 2026 ⏱ 24 min 👁 17 📄 expert opinion 🧭 2026-11-10
Available in: English (current) Français

Keywords

Operational TechnologyAI SecurityVolt TyphoonWater InfrastructureCyber Defense

Summary

In this keynote at the SANS AI Cybersecurity Summit, Joshua Corman, a cybersecurity expert with extensive experience in critical infrastructure protection, delivers a provocative talk warning that AI adoption in operational technology (OT) is a dangerous gamble. He contrasts the reliability of physical infrastructure (steel, concrete) with the undependability of digital systems, urging humility in the cybersecurity community. Corman explains that OT devices like PLCs were never designed for internet exposure and are particularly vulnerable to state-sponsored attacks. He highlights the Volt Typhoon intrusions into US water facilities, framing them as a military campaign aimed at undermining public support during a potential Taiwan conflict. He introduces the “water hammer” analogy to illustrate cascading failures when critical infrastructure is disrupted, and stresses that denial-of-service on OT systems is equivalent to physical destruction. Corman critiques the community’s over-reliance on IT-centric best practices and proposes a pragmatic matrix for evaluating controls: keep, discard, or adapt. He describes his own initiatives, including cybermedsummit.org and UnDisruptable27, which focus on hardening hospital towns and community water systems, and reveals that a national exercise with top insurers exposed severe gaps in OT incident response capacity. He argues that AI should not be introduced into fragile OT environments without careful consideration and instead advocates for reducing dependence, adding physical engineering, or going offline. He closes by announcing his upcoming “Fragile Foundations” report to Congress and calls for engaged, humble action from cybersecurity professionals.

235 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video’s primary value lies in its shift of perspective from IT-centric cybersecurity to the often-overlooked OT and critical infrastructure sector. Corman presents concrete, declassified examples (Volt Typhoon, Iranian reconnaissance) and draws on his personal experience in government and industry, which lends authenticity to his warnings. His argumentation is forceful and coherent, leveraging analogies (water hammer, Jenga towers) to make technical risks accessible. However, his claims are largely anecdotal and rely on his own authority rather than documented citations; he does not provide data or footnotes for many assertions, such as the number of OT incident responders (120) or the details of the insurer exercise. The argument about AI being an additional risk rather than a solution is well reasoned, but it is an opinion piece rather than a systematic evaluation.

Scientific Rigor, Source Quality, Title Accuracy

Corman references several concrete sources, including declassified intelligence reports on Volt Typhoon, his own publications (Rugged Software Manifesto, Fragile Foundations), and his exercise with insurers. He also quotes a 9/11 commission moment to emphasize accountability. However, he does not provide direct citations or links to these sources during the talk; the only link in the description is to the SANS summit registration. This limits the verifiability of his claims. The title is highly appropriate and memorable, distinctly capturing the tension between AI and OT fragility. No comments were provided for analysis.

237 words

Title / Content Match

The title precisely captures the video's core argument: AI introduced into OT environments acts like a bull in a china shop, causing potentially catastrophic consequences. The presentation directly addresses this metaphor.

Quality & Reliability

7/10

The speaker is a recognized expert in critical infrastructure security with direct experience in government and industry. He references declassified incidents (e.g., Volt Typhoon) and describes his own studies and exercises, but the presentation is an opinion keynote rather than a peer-reviewed analysis. Specific claims about adversary capabilities are based on declassified information but not independently verified in this talk.

Key Moments

Cited Sources

Contribution & Novelties

Corman brings a fresh and urgent perspective by highlighting that AI discussions in cybersecurity are overwhelmingly focused on IT, while OT and critical infrastructure—the systems that sustain life—are critically underserved. He proposes a nuanced evaluation framework for existing cybersecurity controls, urging professionals to discard beliefs that are no longer valid, and argues for a pragmatic approach that goes beyond adding AI. His experiences (e.g., the insurer simulation, Operation Warp Speed) offer real-world insights that are rarely discussed in mainstream AI conferences.

Pour aller plus loin :

160 words

Radar Profile

The radar profile shows high scores in information quantity and technical depth but moderate reliability, reflecting a talk rich in expert insights but lacking formal citations. The emphasis on practical experience over peer-reviewed evidence is evident.

Reliability 6/10