AI in OT: Why Humans Stay in the Loop | Jori VanAntwerp | Ep 123

Work with Aaron: https://protectitallpod.com/work/ The book: https://protectitallpod.com/go/book/yt This episode: https://protectitallpod.com/ep123/ The OT Security Starter Kit: https://protectitallpod.com/go/kit/yt AI is a consensus engine, and consensus on the internet is frequently wrong. Jori VanAntwerp, CEO and founder of EmberOT, is back for the most honest AI conversation the show has had: why human-in-the-loop is non-negotiable, and why the junior hires you skip today are the seniors you won't have in ten years. Host Aaron Crow and Jori go deep on sandboxing AI in real workflows, AI-found vulnerabilities that don't matter in OT, teaching AI your voice with markdown primers, the build-vs-buy MVP trap, and why an operator flipping to manual is still the save of last resort. 🎧 Episode page & full transcript: https://protectitallpod.com/ep123/ 🧰 Free OT Security Starter Kit: https://protectitallpod.com/go/kit/yt 📘 The book (Kindle, paperback, hardcover): https://protectitallpod.com/go/book/yt 🤝 Sponsor the show: https://protectitallpod.com/go/sponso... Key Learnings: • AI is a consensus engine. It aggregates what the internet believes, which is often wrong, so your foundations decide whether you can catch it. • Human-in-the-loop is not optional. AI shapes junior output to look senior without the understanding underneath. • An attacker who can reach your HMI or PLC does not need your unpatched CVE. Rank OT vulnerabilities accordingly. • Sandbox rule: AI gets a folder, not the keys. Nothing goes in that you would not publish on the internet. • If one person plus AI does the work of five, nobody is training juniors. No juniors today means no seniors in ten years. • The entry-level job for OT cybersecurity is IT. Help desk teaches you how systems and people actually fail. Key Moments: 0:00 You don't need an exploit (cold open) 0:43 Welcome back Jori VanAntwerp of EmberOT 4:15 AI is a consensus engine 5:05 Not ready for full autonomy: human in the loop 12:10 AI-found vulnerabilities vs OT reality 19:10 Primers: teaching AI your voice with markdown 27:45 The junior pipeline problem 34:50 Vibe coding, skills decay, and the publish-it rule 39:40 Build vs buy: the MVP trap 48:50 The submarine principle: modular defense 52:40 Analog fallbacks: Crash Override and manual mode 59:20 Experience vs time: the hammer story 1:01:00 Breaking in: the entry-level job for OT is IT About the Guest: Jori VanAntwerp is a two-time cybersecurity founder who has spent over two decades helping industrial and IT/OT organizations defend critical infrastructure, with leadership roles at McAfee, FireEye, CrowdStrike, Dragos, and Gravwell before founding SynSaber and now EmberOT, where he focuses on visibility, security, and risk quantification for OT environments. Connect with Jori:   / jvanantwerp   EmberOT: https://emberot.com | Free OT PCAP Analyzer: https://www.emberot.com/ot-pcap-analy... Jori's earlier episodes: Ep 49    • S4 Conference Highlights: Where IT Meets O...   and Ep 29    • IT & OT Cybersecurity for Power Plants wit...   Protect It All: All Things Cybersecurity, from IT to OT Hosted by Aaron Crow (about: https://protectitallpod.com/#about) Website: https://protectitallpod.com Blog: https://protectitallpod.com/blog/ Apple Podcasts: https://podcasts.apple.com/us/podcast... Spotify: https://open.spotify.com/show/1Vvi0eu... X:   / protectitall   Facebook:   / protectitallpodcast   Be a guest or suggest a topic: info@protectitall.co #aicybersecurity #otsecurity #artificialintelligence #cybersecurity #protectitall