EP31

Industrial AI vs GenAI: What Really Works on the Factory Floor?

With Nikita Golovko, Portfolio AI Architect, Siemens

Key takeaways

  • Predictability beats hype on the shop floor: classical ML models with predictable behavior are still the backbone of industrial AI, while generative AI plays supporting roles like documentation and synthetic data generation.
  • Environment matters: AI that works in the lab will break in production if you ignore unstable conditions, noisy data and real-world user behavior.
  • Translators are critical: you need people who can bridge data science metrics and operational KPIs — this is the real role of an AI architect in industry.
Read the full write-up →

Show notes

Welcome to the Impact of AI: Explored podcast. In this episode
James and Gerjon have a awesome guest: Dr. Nikita Golovko - Certified ISAQB CPSA-F Architect/Cybersecurity expert & PSSE (ISC2 CC) (https://www.linkedin.com/in/dr-nikita-golovko/)

In this conversation, we discuss the challenges and nuances of implementing AI in industrial settings. We look at the importance of bridging the gap between AI technology and practical application on the factory floor, highlighting the need for translators who can communicate between data science and operational teams. The discussion also covers the transition from lab-based AI to real-world applications, the significance of team structure in successful AI projects, and the responsibilities of engineers in ensuring AI systems are safe, transparent, and sustainable. We share insights on generative AI, technical debt, and the future of AI in industrial automation, advocating for a collaborative approach that includes human oversight in decision-making processes.

Takeaways
* Finding the right talent to bridge AI and operational gaps is crucial.
* AI solutions often fail when transitioning from lab to real-world environments.
* Technical debt can be identified by signals like fear of touching code.
* Generative AI is not yet suitable for industrial applications due to unpredictability.
* AI should enhance human decision-making, not replace it.
* Team structure significantly impacts the success of AI projects.
* Sustainability in AI means retraining models with new teams over time.
* Transparency and explainability are essential for AI systems.
* AI can assist in coding but should not replace foundational knowledge.
* Documenting decisions is key to managing technical debt.

Chapters

  1. 00:00AI agents in industrial environments
  2. 01:33introduction
  3. 02:18Listener question: Bridging the Gap in Industrial AI Hiring
  4. 06:47Challenges of Deploying AI in Real-World Environments
  5. 09:38Technical Debt and Its Early Warning Signs
  6. 12:18Generative AI vs. Classical Machine Learning in Industry
  7. 14:51The Role of Predictability in AI Models
  8. 17:29Local vs. Centralized AI Models in Production
  9. 20:14Architectural Challenges in AI Implementation
  10. 22:55Sustainability in AI System Design
  11. 25:52The Future of AI in Industrial Applications
  12. 26:54The Role of AI in Industrial Automation
  13. 32:00AI as an Advisor: The Human Element
  14. 36:39AI-Assisted Coding: Benefits and Risks
  15. 40:08Responsibility and Ethics in AI Development

About our guest

Dr. Nikita Golovko

Portfolio AI Architect, Siemens

Certified iSAQB architect and cybersecurity expert. Talked about industrial AI versus GenAI on the factory floor.

LinkedIn → All appearances →

Join the conversation

What did you think? Tell us on LinkedIn, send us a note, or suggest someone we should talk to.