Can artificial intelligence predict and prevent diseases before clinical symptoms appear?
Will smart factories balance automation with safe human–robot collaboration? How close are
we to AI-managed air travel and single-pilot cockpits? These questions guided the fifth session
of the global course “Understanding AI and Robotics: Their Multi-Dimensional and Multi-Spatial
Implications for the Public and Private Sector”, held on 19 June 2025 under the title “Smart
Production in the Connected World.”
Spanning the fields of medical technology, industrial manufacturing, product engineering, and
aviation, the 5 th lecture brought together three distinguished experts: Dipl.-Ing. David Ljuhar,
founder of BRAINCON Technologies, Captain Rehan van Tonder, CEO of Shift Aviation
Solutions and Boeing 777 pilot, and Dr. Sebastian Schlund, Head of the Institute of Production
Engineering at TU Wien.
Framed by topics such as Industry 4.0, AI-powered logistics, healthcare prediction models, and
safe aviation, the session offered participants both practical insights and critical reflections on
what smart production truly demands in an increasingly connected world.

AI for Healthcare and Infection Control: Davul Ljuhar’s Perspective
Dipl. Ing. Davul Ljuhar delivered a lecture on the applications of AI in healthcare diagnostics and
infection control. Drawing from decades of professional experience in high-tech industries and
medical engineering, Ljuhar detailed two main areas of his expertise and work: AI in predictive
diagnostics of bone-related diseases, and AI in the control and prevention of hospital-acquired
infections.
Ljuhar began his presentation by reflecting on his initial encounter with AI, that took place in
1987, during his research on the use of expert systems in mechanical design. He later founded
BRAINCON Technologies and redirected his focus to the assessment of bone diseases,
specifically osteoporosis.
In collaboration with French researchers, Ljuhar’s team shifted from bone density to bone
quality as the primary indicator of bone health. They used fractal dimension analysis to assess
bone structure, establishing a methodology to evaluate the quality of bone tissue. This required
large datasets and clearly defined ground truth standards.
Between 2012 and 2018, his team analysed a 10-year retrospective dataset from San Francisco.
It comprised 7.4 terabytes of data from 7,000 patients. From this data, they isolated 2,000
baseline patients with no visible symptoms. Of these, 1,000 later developed osteoarthritis.

Ljuhar’s analysis confirmed that predispositions to disease were visible in baseline data,
proving that AI could detect early structural anomalies well before clinical symptoms appeared.
Building on this foundation, his team expanded AI applications to diagnose rheumatoid arthritis
and osteoporosis. Ljuhar noted parallels in current mammography technology, where AI
systems can detect precancerous lesions up to three years in advance.

AI in Infection Prediction and Control
From 2016 onward, Ljuhar shifted focus to nosocomial (hospital-acquired) infections. He
emphasized the global urgency of this issue, citing an estimated 36 million annual deaths due to
such infections, with a projected increase to 170 million by 2050. Ljuhar identified a need for
systems that combine software and hardware to support hospitals in disinfection and outbreak
prevention.
A significant challenge lay in the diversity of available data, which often lacked compatibility.
Ljuhar illustrated this with an analogy: translating technical data from German to Chinese and
back again resulted in inconsistent information. His team worked to harmonize and structure
this data by identifying correlated data clusters. They defined four initial data groups to analyse
environmental factors such as humidity, temperature, and chemical concentration (e.g.,
hydrogen peroxide) in hospital settings.
Using convolutional neural networks (CNNs), his team developed algorithms capable of
predicting effective germ inactivation strategies. Ljuhar highlighted the potential of
reinforcement learning and cognitive systems in future infection control, anticipating AI-
supported systems that could adapt and improve based on environmental data and past
performance.
He referenced a current case in Kenya, where hospital infection rates contribute to 32% infant
mortality among newborns. According to Ljuhar, such AI systems could significantly lower
infection-related deaths and costs, which are expected to reach over $200 billion by 2050. In
Austria alone, 140,000 cases of nosocomial infection are reported annually.
Ljuhar concluded by stating that his company’s pilot systems would be tested in hospitals
starting later this year. He stressed that AI could provide meaningful support in global
healthcare management and serve as predictive tools for both chronic diseases and public
health crises.

AI in the Aviation Sector: Rehan Van Tonder’s Perspective
Captain Rehan Van Tonder, CEO of Shift Aviation Solutions Group, a former Boeing 777 captain
and instructor, introduced the foundational structure required for operating an airline,
emphasizing five essential components: physical hardware, human software, an operational

base, financial support, and legal authorization. He underscored that the current aviation model
faces persistent inefficiencies. These include long wait times for check-in, boarding, and
immigration processes. According to Van Tonder, AI can directly address these inefficiencies
through automation and predictive planning.
Van Tonder highlighted how third-party services are set to revolutionize passenger logistics. AI-
powered drones could transport baggage directly from a passenger’s home to the aircraft.
Biometric verification systems, already in use for passport control, are likely to eliminate the
need for physical boarding passes. Wearable devices like smartwatches will soon enable
biometric identification from check-in to boarding. These advancements will streamline the
entire passenger journey.

AI in Flight Operations and Safety
Rehan Van Tonder stressed the vast volume of operational data transferred between aircraft
and systems such as flight planning, maintenance, and checklist management. He emphasized
that AI could translate this data into actionable insights, enhancing cognitive support for human
operators. AI will reduce pilot errors by filtering and prioritizing information in high-pressure
scenarios.
He presented a model for future flight decks, which may involve single-pilot operations. AI
systems would take on procedural and emergency functions, leaving the pilot to supervise and
intervene in exceptional cases. In the long term, pilotless cockpits might become feasible,
although safety protocols and insurance frameworks remain barriers.

Economic Implications and Human Roles
Responding to a question about operational costs, Van Tonder explained that pilots contribute
roughly 10% to an airline’s overall expenditure, with hardware, software, and fuel costs each
constituting about a third. Automation could reduce the number of pilots needed on long-haul
flights, generating cost savings without compromising safety. Cabin crew responsibilities will
also evolve, with AI-enabled robots managing non-critical tasks like food and beverage service,
while humans focus on safety and security.
Van Tonder anticipated future developments in AI-human interaction. Wearable cognitive
augmentation devices could monitor pilot stress levels in real time. Ethical questions arise: who
owns the data collected from a human’s brain activity? Will cognitive ability disparities lead to
inequality in interfacing with advanced systems?
He warned of potential consequences if AI systems malfunction or behave unpredictably. AI
oversight mechanisms must ensure fallback options, such as remote piloting, remain viable.

AI Solutions for Airport Efficiency
Van Tonder presented innovations developed by his company, including AI software that
sequences aircraft arrivals to minimize holding patterns and reduce delays. This approach could
eliminate the need for costly airport expansions. He cited the example of Montenegro, where his
company’s software projected annual savings of over 70 million euros.
These solutions aim to optimize existing infrastructure, reduce CO2 emissions, and increase the
efficiency of resource use across the aviation ecosystem.

Health Risks and Radiation Exposure
Van Tonder also addressed concerns about radiation exposure in aviation. He cited peer-
reviewed research showing varied outcomes but confirmed that exposure does exist, especially
for flight crews. Airlines monitor solar activity and adjust flight paths to mitigate risk. The
implementation of AI may allow real-time radiation monitoring and further reduce health risks.
Van Tonder concluded by advocating for AI as a tool to increase safety, efficiency, and
sustainability in aviation. He emphasized that while AI cannot yet replace human judgment in
full, it can augment decision-making, reduce human error, and streamline operations. However,
ethical, legal, and infrastructural challenges must be addressed.

The Rise of Humanoid Robotics in Industry: Dr. Sebastian Schlund’s Perspective
Dr. Schlund brings decades of expertise in industrial engineering and human-machine
interaction. He has held academic and leadership roles at the University of Stuttgart, Fraunhofer
IAO, and currently leads applied research at TU Wien. His current work focuses on the interface
between robotics, production processes, and human-centered system design.
Sebastian Schlund began his lecture by outlining the historical trajectory of humanoid
robotics. Early milestones included Electro, a remote-controlled robot built in 1939 by
Westinghouse, and Japan’s WABOT-1 in 1973, capable of basic walking and gripping.
These early efforts laid the groundwork for today’s more sophisticated humanoid
systems, which now benefit from modern AI technologies.
Recent advancements by companies such as Tesla (Optimus), Boston Dynamics (Atlas),
and Agility Robotics (Digit) have shown the potential of humanoid robots in performing
simple, structured tasks. However, Dr. Schlund emphasized that these demonstrations
remain mostly confined to controlled settings and are not yet robust enough for
widespread deployment.

Simulation, AI, and Hardware Integration
Dr. Schlund highlighted the critical role of advanced AI and simulation platforms in recent
progress. Tools developed by NVIDIA and OpenAI, for instance, allow developers to simulate
robotic movement and learning in virtual environments before transferring these capabilities to
physical robots. This simulation-to-reality pipeline has accelerated the development process
and reduced costs.
Still, challenges persist. The physical limitations of hardware—especially robotic hands or end
effectors—represent a major bottleneck. State-of-the-art five-fingered robotic hands can cost
upwards of €20,000 to €30,000 per unit. Despite their mechanical sophistication, these tools
still lack the flexibility and responsiveness of human hands, particularly when interacting with
soft or deformable materials.

Human-Robot Collaboration and Future Trajectory
Dr. Schlund underscored the importance of designing collaborative robotics systems. He argued
that humanoid robots should not aim to replace human labor but rather to support workers in
environments where ergonomic stress or safety risks are high. In his view, true productivity
gains will emerge not from automation alone but from balanced cooperation between humans
and machines.
Looking forward, Dr. Schlund projected a realistic timeframe of about 10 years before humanoid
robots could see large-scale deployment in industry. This will require not only technological
refinement but also adaptation of workflows, regulatory frameworks, and workforce training.

(Author Kamila Bogdanova is a Research Assistant at the Institute for the Danube Region and Central Europe (IDM) in Vienna, and Information Officer at the International Institute for Middle East and Balkan Studies (IFIMES).