Trends in industrial artificial intelligence were part of ARC Leadership Forum, 2026 discussions, including how AI software integration into automation software helps autonomous operations.
While artificial intelligence (AI) has been used in automation for more than 20 years, automation and AI are moving to autonomous operations. With simulation and AI combined, automation is moving into a different space, said Jim Chappell, Aveva, vice president, global head of AI, in a discussion with Control Engineering at the 2026 ARC Leadership Forum by ARC Advisory Group, Feb. 9-12, 2026, Orlando, Florida. Theme of the 30th Annual ARC Industry Leadership Forum event is “How AI Is Driving the Future of Industrial Operations and Supply Chain.”
How AI helps industrial autonomy in industrial operations
AI can help automation’s progression toward more autonomous operations with the ability to apply pseudo-sensors where sensors are limited or hard to apply. Beyond that, it can use cutting-edge AI techniques to predict what the values should be so that operators can make control decisions. Taken further in a closed loop, the system can operate itself under human supervision, Chappell said.
As AI model training advances, more AI is being applied in industrial applications. At present, physics-based AI applications often include humans in the loop or on the loop, monitoring AI suggestions, Chappell said. High quality, simulated data is often used to train advanced AI models to provide reliable closed-loop operations for quality production with less downtime.
Systematically, reinforcement learning models are being applied to industrial AI and are producing autonomy for some applications at levels 3 or 4 based on the SAE J3016 standard for driving automation systems. It’s not there yet for Level 5, Chappell said.
Industrial AI use for simulation
A dynamic simulation tool can be described as a white box, Chappell said, a visible way to produce a simulation and see the physics-based algorithms used to produce the results. AI can be described as a black box, by using data-driven AI techniques to produce results without the clear insight into how the results were determined. An optimal mix is often by leveraging AI in the loop with simulation, a gray-box application. Physics algorithms are leveraged where they perform best, and AI is intermixed and used where it performs best. Often, AI is being applied to speed up and deploy simulation more quickly, Chappell said. In addition, applications can use simulated data to train AI with both physical and pseudo sensors for model improvement. AI-enhanced simulation can be used in cloud applications or on the edge to improve efficiency and optimize setpoints in a control system. Applications include chemical, food and beverage and oil and gas process automation.
Physics informed neural networks (PINN) use physics as part of the loop, allowing tighter integration of the physical world.
Industrial AI, layers, with governance, trust
As automation tools, AI included, become more accepted, they tend to become integrated into applications, less visible and more usable. Types of AI are being layered into applications, Chappell said, such as generative AI and language models, with the idea of going after the biggest fish. End-to-end AI governance is needed, he said, with office, commercial and internal operational systems.
Aveva has 21 AI-infused products, many usable with self-service or minimal training. Internally, there’s a main working group looking at interoperability of tools at the data level and at connections among software, including streamlined user experiences that leverage AI to guide users.
Adding use, trust with industrial AI integration
What are recommendations for those interested in industrial AI?
Integrate industrial AI where it makes sense, Chappell said. AI is used across the entire industrial life cycle, from engineering design to operations and optimization with design simulation, to create and improve controls and provide 3D visualization. People are building trust as they use commercially available AI tools for personal requests on how to do and improve things.
Connected platforms will pull AI together across applications, Chappell said.
Mark T. Hoske is editor-in-chief, Control Engineering, WTWH Media, [email protected].
Keywords
Industrial AI software, AI integration, industrial AI platforms
Consider this
AI is integrating into industrial automation software. Are you getting what you need?
You also might like
Control Engineering provides more about AI and machine learning
https://www.controleng.com/ai-and-machine-learning
Aveva explains more about AI software in this Control Engineering article
Responsible AI for industry: At scale with results
https://www.controleng.com/responsible-ai-for-industry-at-scale-with-results
ARC Advisory Group provides more about the ARC Industry Leadership Forum
https://www.arcweb.com/events/arc-industry-leadership-forum-orlando
