Editor’s Note: Artificial intelligence is beginning to change how physical work is performed and measured. One of the most visible developments is the use of powered exoskeletons in industries such as healthcare, logistics, transportation, and manufacturing.
The Exia, developed by German Bionic, is designed to assist workers with lifting and walking tasks. Worn by people rather than operating autonomously, the device supplements human movement to reduce strain and improve safety. Each suit connects to a cloud platform that tracks key metrics such as safety scores, fatigue, and device utilization. The data helps employers identify risks, adjust workloads, and evaluate the impact of wearable robotics on ergonomics and productivity.
In this interview, chief product officer Norma Steller explains how AI enables exoskeletons to adapt to individual users, what challenges come with training models for variable industrial settings, and how companies are approaching data governance and privacy as wearables become part of everyday work. German Bionic has partnered with KULR Technology to introduce this novel technology to supply chain operations in the United States and Canada
DSG: How is artificial intelligence reshaping the design and functionality of modern industrial exoskeletons? What kinds of sensing, learning, or adaptive algorithms are emerging as most critical for real-time movement assistance?
Steller: Artificial intelligence is fundamentally transforming industrial exoskeletons by enabling the devices to adapt to the unique movement patterns of each individual user. Unlike machines that operate with fixed settings, exoskeletons must accommodate the highly personal and varied ways people move. Since users cannot always clearly express the nuanced support they require at different points in their movements, AI becomes essential in interpreting these subtleties and closing this communication gap.
Emerging AI algorithms focus on real-time sensing and continuous learning from actual usage data. These types of adaptive systems analyze the subtle differences in motion and adjust assistance dynamically without requiring the user to manually input preferences. This allows the exoskeleton to evolve alongside the user’s habits and physical conditions, providing personalized support that feels natural and intuitive. Such intelligent adaptation is essential for improving comfort, reducing fatigue, and enhancing overall effectiveness in industrial settings.
DSG: Many next generation exosuits are described as “intelligent” or “autonomous.” From a technical standpoint, what distinguishes a true AI-enabled system from one that’s simply sensor-driven or pre-programmed?
Steller: The key distinction lies in the system’s ability to learn and adapt autonomously. While sensor-driven or pre-programmed exoskeletons respond to fixed inputs or simple triggers, true AI-enabled systems continuously analyze user behavior and environmental conditions to refine their responses over time. This learning capability allows the exoskeleton to tailor its assistance to the specific needs and movement patterns of each individual user.
Without this adaptive learning, exoskeletons can often feel rigid or uncomfortable because they cannot adjust to the nuances of human motion. In contrast, AI-powered exoskeletons feel more like an extension of the body, providing support that evolves naturally with the user.
DSG: How are data and connectivity changing the value proposition of wearable robotics? What types of analytics, cloud integration, or digital-twin concepts are helping employers measure ROI and safety outcomes?
Steller: Data and connectivity are revolutionizing wearable robotics by making the benefits of exoskeletons measurable and actionable. By collecting detailed information on how workers move and exert themselves, these systems provide valuable insights into the physical demands placed on employees. This data helps employers understand the preventative health benefits of exoskeleton use, encouraging adoption by demonstrating clear safety and productivity improvements.
Advanced analytics and cloud integration enable decisionmakers to visualize and quantify human labor in ways never before possible. For example, digital twins allow for precise simulation and optimization of workflows. This visibility helps companies identify specific tasks or areas where support is most needed, leading to targeted interventions that save costs and enhance worker safety and well-being.
Moreover, aggregated data supports regulatory bodies by providing real-world evidence to inform and evolve workplace safety standards, fostering safer industrial environments overall.
DSG: Industrial environments are notoriously variable. What are the biggest challenges in training AI models for exoskeletons that must perform reliably across warehouses, construction sites, or manufacturing lines?
Steller: One of the greatest challenges is building a scalable and efficient infrastructure that balances computation between the exoskeleton device and cloud services. Industrial exoskeletons often operate in environments without reliable standby power or continuous internet connectivity, so developers must decide which AI processes run locally on the device and which run remotely in the cloud.
Additionally, the AI models need to be compact and efficient enough to execute on limited hardware without compromising responsiveness or accuracy. Achieving this balance is critical for ensuring the exoskeleton can adapt in real-time across diverse and unpredictable settings like warehouses, construction sites, and manufacturing floors.
A further – and often underestimated – challenge is securing the breadth and quality of real-world data needed to train AI models that remain robust across such variable environments. Exoskeleton performance cannot be perfected in controlled lab conditions alone; it depends on understanding how real people move during real tasks under real constraints. Here, German Bionic’s long-standing deployment of fully connected exoskeletons has created a uniquely rich foundation: billions of motion data points gathered from diverse industries, job types, and environmental conditions. This accumulated knowledge provides an empirical basis for developing AI models that generalize reliably, capturing the subtle variations in human movement and task execution that occur across warehouses, construction sites, and manufacturing lines. It is this depth of real-world experience that enables the next generation of exoskeletons – such as the Exia – to perform consistently in environments where variability is the norm, not the exception.
DSG: Where does machine learning make the most impact — motion prediction, adaptive torque control, user fatigue detection, or safety overrides? And how do developers ensure transparency and trust when the system makes real-time adjustments?
Steller: Machine learning’s greatest impact lies in helping the exoskeleton understand and react to individual movement patterns, enabling adaptive control that supports the user’s natural motion. Importantly, real-time adjustments are subtle and occur behind the scenes, so users do not feel sudden or intrusive changes during movement.
The user always maintains full control; the exoskeleton functions more like an advanced assistive tool rather than an autonomous robot. This approach is comparable to assisted driving systems in cars, which help reduce complexity without taking over control completely.
Transparency and trust are ensured by designing the system to operate predictably and by communicating clearly that the exoskeleton supports rather than replaces the user’s own actions.
DSG: How are AI-driven exoskeletons tested and validated for safety and ergonomic impact? What emerging standards or testing protocols are guiding this space?
Steller: As a European company, German Bionic complies with stringent European Union regulations such as the Data Act and the AI Act, which emphasize privacy, transparency, and safety. The company collaborates with independent safety experts like TÜV to rigorously test exoskeletons across all safety dimensions.
Although there are currently no dedicated testing protocols specifically for exoskeletons, manufacturers typically follow established guidelines for connected wearables and machinery. This approach ensures that AI-driven exoskeletons meet exacting standards for both user safety and ergonomic effectiveness.
DSG: Many companies see exoskeletons as part of a broader “human-plus-machine” productivity model. How should organizations think about integrating these tools alongside robotics, automation, and AI-powered workflow systems?
Exoskeletons represent a unique synergy between human flexibility and machine strength. They are especially valuable in roles requiring direct human interaction, rapid adaptability, and operation in unstructured environments—such as elderly care, emergency response, logistics, and construction.
Organizations should view exoskeletons as complementary to robotics and automation, enhancing human capabilities rather than replacing them. Integrating these wearables with AI-powered workflow systems can optimize productivity by leveraging the best of both human intelligence and machine endurance.
DSG: With data now streaming from wearables, what are the implications for worker privacy and data governance? How can companies balance performance insights with ethical and legal considerations?
Steller: German Bionic adheres to the principle of privacy by design, meaning data collection is limited strictly to what is necessary. The exoskeleton can function and improve its AI models without requiring any personal data from the user.
This careful approach ensures that companies gain valuable performance insights while respecting workers’ privacy and complying with legal frameworks. Transparency about data use and robust governance policies are essential to maintaining trust.
DSG: AI hardware advances — from edge chips to low-latency connectivity — are evolving rapidly. How might these innovations change what’s possible for real-time biomechanical assistance in the next three years?
Steller: While hardware improvements like edge computing and faster connectivity are important, the real breakthrough comes from developing efficient algorithms that require minimal processing power. German Bionic’s technology already runs sophisticated AI models directly on the exoskeleton in offline mode, without needing expensive specialized CPUs or GPUs.
This capability allows for more seamless integration of exoskeletons into daily work life, making them more accessible and practical. Over the next few years, such innovations will help bring these devices into the mainstream by improving usability and reducing costs.
DSG: Looking ahead, what breakthroughs could make AI-powered exoskeletons truly mainstream? Will success depend more on AI sophistication, affordability, user experience, or seamless integration into industrial ecosystems?
Steller: The future success of AI-powered exoskeletons depends on a combination of factors, with user experience being the most critical. For these devices to become truly mainstream, they must feel like a natural part of the body—like a second skeleton.
Developers will continue to strive to reduce complexity and friction with every new iteration, ensuring the exoskeletons are comfortable, intuitive, and easy to use. Alongside AI sophistication, affordability, and ecosystem integration, enhancing user experience will determine widespread adoption.
Don’t miss any content from Distribution Strategy Group. Join our list.
