# AI in OT Should be Cheaper Than in IT Source: https://www.youtube.com/watch?v=CPlLG3yHkrg Recap page: https://rapidrecap.app/video/CPlLG3yHkrg Generated: 2025-09-26T15:36:17.551+00:00 --- ## Quick Overview AI in OT cybersecurity is complex and costly due to the age and diversity of systems, but applying a balanced approach and leveraging AI effectively can enhance security posture, though it requires careful consideration of network architecture and skilled personnel. **Key Points:** - AI's effectiveness in OT cybersecurity is optimized by human intelligence and understanding of the specific network architecture. - Careful consideration of OT network architecture is crucial when selecting and implementing AI cybersecurity solutions. - Organizations can build more resilient and secure OT environments by implementing a balanced approach to AI. - OT cybersecurity requires a multi-faceted approach, combining AI strengths with the expertise of cyber specialists and OT operators. - Implementing AI in simpler OT environments is less demanding, focusing on unified networks and requiring less data processing and analysis compared to multi-layered architectures. - AI in complex OT environments requires more effort, demanding understanding of unique communication patterns and potential vulnerabilities across multiple layers. - Pricing for AI deployment in OT environments varies significantly, with simple OT environments costing $50k-$200k and complex OT environments costing $500k-$2M+. ![Screenshot at 00:00: The video begins with the AIxCC logo and the title "AIxCC Stage at DEF CON 33", setting the context for a discussion on AI in cybersecurity.](https://ss.rapidrecap.app/screens/CPlLG3yHkrg/00-00-00.png) **Context:** The video discusses the challenges and strategies for implementing AI in Operational Technology (OT) cybersecurity, highlighting the differences between IT and OT environments. It emphasizes that OT systems are often older, more diverse, and less frequently updated than IT systems, making them more vulnerable. The presentation touches upon the Purdue Model for network segmentation and the concept of Zero Trust, suggesting that a more tailored approach is needed for OT security. The speaker uses pop culture references from Star Wars and Terminator to illustrate concepts. ## Detailed Analysis The presentation "AI in OT Should be Cheaper Than in IT" explores the complexities and costs associated with implementing AI for cybersecurity in Operational Technology (OT) environments. Unlike IT environments, OT systems are often older, built with legacy technology, and have unique operational requirements, leading to higher implementation costs and specialized needs for AI solutions. The speaker, Daryl Haegley, Technical Director of Air Force & Space Force Control Systems Cyber Resiliency, highlights that AI in OT environments works harder due to the need to understand unique communication patterns and potential vulnerabilities across multiple layers of the Purdue Model. He contrasts the 'simple' and 'complex' OT environments, noting that simpler setups are less demanding for AI implementation, while complex ones require more effort and higher costs ($500k-$2M+) due to advanced cybersecurity measures and custom API development. The presentation also touches upon the concept of Zero Trust, suggesting that a shift from perimeter-based security to a more granular approach is necessary. The speaker stresses the importance of a balanced approach, combining AI's strengths with human expertise and careful consideration of OT network architecture. He also touches on the challenges of pricing AI solutions for OT, noting that the costs are significantly higher than for IT due to the specialized nature of OT systems and the need for custom solutions. The talk concludes with a call to action for better contract requirements and performance considerations for AI in OT, emphasizing the need for specialized training and expertise. ### AI in OT vs. IT - AI is more complex and costly in OT due to older systems, diverse architectures, and unique operational requirements. ### Purdue Model and Zero Trust - The Purdue Model's hierarchical framework is discussed in relation to AI's role in monitoring and securing OT networks. Zero Trust principles are highlighted as a necessary shift from perimeter security. ### Cost of AI Deployment - Simple OT environments cost $50k-$200k, while complex environments cost $500k-$2M+, reflecting implementation time, integration points, cybersecurity measures, and training. ### Challenges in OT Security - OT networks have unique vulnerabilities and require specialized approaches to AI implementation, with a need for skilled personnel and tailored solutions. ### Balancing AI and Human Expertise - The effectiveness of AI is optimized when guided by human intelligence and understanding of OT networks, requiring a multi-faceted approach. ### Future Considerations - The need to improve contract requirements and performance for AI in OT is emphasized, along with the importance of training and expertise. ![Screenshot at 00:00: The AIxCC logo and title screen for the "AI Cyber Challenge" stage at DEF CON 33.](https://ss.rapidrecap.app/screens/CPlLG3yHkrg/00-00-00.png) ![Screenshot at 00:04: A montage of glitchy code and data streams, symbolizing the digital and potentially vulnerable nature of OT systems.](https://ss.rapidrecap.app/screens/CPlLG3yHkrg/00-00-04.png) ![Screenshot at 00:06: A text overlay reads "AI IN OT SHOULD BE CHEAPER THAN IN IT", highlighting the central theme of the presentation.](https://ss.rapidrecap.app/screens/CPlLG3yHkrg/00-00-06.png) ![Screenshot at 00:16: The presenter, Daryl Haegley, is shown, indicating the start of his presentation.](https://ss.rapidrecap.app/screens/CPlLG3yHkrg/00-00-16.png) ![Screenshot at 00:17: A slide displays the "CYBER RESILIENCY OFFICE FOR CONTROL SYSTEMS" \(CROCS\) logo, along with unclassified status and date.](https://ss.rapidrecap.app/screens/CPlLG3yHkrg/00-00-17.png) ![Screenshot at 00:51: A diagram illustrating "100% Military Missions Depend on Installation Operational Technology", detailing various base systems.](https://ss.rapidrecap.app/screens/CPlLG3yHkrg/00-00-51.png) ![Screenshot at 01:03: The speaker gestures towards a slide explaining the Purdue Model, a hierarchical framework for OT network segmentation.](https://ss.rapidrecap.app/screens/CPlLG3yHkrg/00-01-03.png) ![Screenshot at 02:41: A bar graph comparing the number of devices in "SECURITY" vs. "INFO SYS" vs. "CONTROL SYSTEMS / OT", showing a significant disparity.](https://ss.rapidrecap.app/screens/CPlLG3yHkrg/00-02-41.png) ![Screenshot at 04:31: A slide comparing the "Purdue Model" to a "Flat Architecture", illustrating the complexity of layered networks.](https://ss.rapidrecap.app/screens/CPlLG3yHkrg/00-04-31.png) ![Screenshot at 07:35: A diagram titled "Focused Defender in a Streamlined OT Network" shows AI monitoring a simplified network structure with various devices connected to a switch and AI monitor.",](https://ss.rapidrecap.app/screens/CPlLG3yHkrg/00-07-35.png)