Industry panel at TSIS 2026 examines the future of digital twins in defense training

By Cassidy Mills, Senior Staff Writer

August 5, 2026

As digital twins, synthetic environments and artificial intelligence (AI) continue to reshape defense technology, industry leaders gathered at the 2026 Training & Simulation Industry Symposium (TSIS) on June 17 to discuss how these innovations are transforming military training, testing and operational readiness.

Presented by the National Training & Simulation Association (NTSA), TSIS is one of the key gatherings for the modeling, simulation and training (MS&T) community. Within Florida, it is second in scale only to the Interservice/Industry Training, Simulation and Education Conference (I/ITSEC).

The symposium serves as an annual forum for representatives from industry and the U.S. military to discuss training and simulation requirements. Officials from the Army, Navy, Marine Corps and Air Force presented updates on current and future procurement needs. Topics included acquisition strategies, timelines and funding priorities.

Titled “Digital Twins, Synthetic Environments, and the Future of Persistent Training,” the session explored how digital twins are evolving beyond traditional simulation tools to become the foundation for adaptive training, AI development and more informed operational decision-making.

Moderated by Alethea Duhon, Ph.D., director of modeling, simulation and training at KBR, the panel featured Pritesh Patel, chief executive officer of Ad Hoc Research Associates; Tim Rothenburger, director of engineering at The DiSTI Corporation; Justin Wright, technical solutions architect at World Wide Technology; and Audrey Zlatkin, research, development, test and evaluation (RDT&E) portfolio manager, digital twins at Design Interactive.

Throughout the discussion, panelists explored the technical, organizational and human challenges involved in implementing digital twins while considering their future role across the defense enterprise.

The conversation opened with what Duhon called one of the industry’s most fundamental questions: What exactly is a digital twin? While the term is widely used across both defense and commercial sectors, the panelists agreed there is still no universally accepted definition.

Patel defined a digital twin as a virtual representation that remains directly connected to a specific physical asset through continuous, two-way communication.

“A simulation can exist independently. It doesn’t require a physical asset to exist,” Patel said. “A digital twin only exists because that specific asset exists.”

He added, without bidirectional communication between the physical and virtual systems, the technology is better described as a “digital shadow” than a true digital twin.

Zlatkin expanded the discussion by emphasizing that digital twins should represent more than equipment alone. As the technology matures, she said organizations are increasingly integrating physiological and behavioral data to better understand the human operator alongside the system itself.

“We’re looking not only at what’s happening with the equipment or the system, but also what’s happening with the human operator,” Zlatkin said. “That includes things like fatigue, trust, workload, cognitive state and overall readiness.”

Turning to implementation, Rothenburger said organizations often underestimate the work required to connect existing infrastructure and prepare data for digital twin applications. While many organizations already possess the necessary information, he said it frequently resides in disconnected systems that require significant integration efforts.

“Once you decide you’re going to build a digital twin, you discover that an enormous amount of time, effort and money has to be invested in cleaning, transforming and connecting data pipelines,” Rothenburger said. “That’s not necessarily the exciting part of building a digital twin, but it’s often the hardest part.”

Wright noted maintaining digital twins throughout their lifecycle presents another challenge. Organizations may invest in building the technology, but he said long-term ownership and maintenance responsibilities are often unclear.

“Everyone wants the benefits of a digital twin, but when it comes to maintaining it over time, updating the data and keeping it synchronized with the real world, it isn’t always clear who owns that responsibility,” Wright said.

The panel also examined what defense can learn from commercial industries already using digital twins. Patel pointed to telecommunications and aviation as examples where operators routinely evaluate multiple courses of action before making mission-critical decisions.

While those concepts can benefit defense, Patel noted military environments introduce far more variables and uncertainty than most commercial applications. Zlatkin added that accounting for changing human performance makes defense applications even more complex, requiring digital twins to model both equipment and operator readiness.

Looking ahead, the conversation shifted to the growing role of AI within digital twin environments. Patel explained many AI systems are developed in controlled environments that fail to reflect operational conditions. He said digital twins provide realistic settings where AI can be tested, refined and validated before deployment.

“A digital twin provides a much more realistic environment where AI can be developed, evaluated and refined before it’s deployed,” Patel said. “You can identify problems early instead of discovering them after deployment.”

Wright said synthetic environments also accelerate AI development by generating labeled training data and recreating rare or hazardous scenarios that would be difficult — or impossible — to capture in real-world operations.

“Synthetic environments can generate labeled training data automatically,” Wright said. “They allow us to create edge cases that would either be dangerous, expensive or nearly impossible to capture in real-world operations.”

As the session concluded, the panelists shared their outlook for the future of digital twins. Patel described AI as driving a shift toward software that continuously adapts using operational data, while Wright encouraged organizations to rethink the purpose of simulation.

“We shouldn’t simply build larger simulations,” Wright said. “We should build smarter simulations. The objective isn’t just to recreate reality. It’s to generate new knowledge, improve decision-making and support better training.”

Zlatkin closed by encouraging organizations to prioritize interoperability from the beginning of their programs rather than attempting to solve integration challenges after systems have already been developed.

“My recommendation is to think about integration requirements early in your programs instead of trying to solve interoperability after systems have already been built,” Zlatkin said.

As defense organizations seek to build more connected and resilient training ecosystems, panelists agreed that success will depend on integrating technology, data and human performance into a unified digital environment.