ERF 2026

Summary of the Workshop

The 13th Workshop on Hybrid Production Systems (HPS) was moderated by George Michalos (Laboratory for Manufacturing Systems and Automation – LMS, University of Patras). He opened the session with a concise overview of previous HPS activities and introduced the workshop’s main thematic pillars. The workshop attracted more than 60 participants from both academia and industry, reflecting strong interest in the field.

The workshop aimed to present current R&D trends in robotics and Artificial Intelligence (AI) for human-centred production systems, with a focus on AI-enabled robot cognition, flexible production through simulation and digital environments, and safe human–robot collaboration. It also addressed workforce upskilling and human-in-the-loop approaches supported by immersive technologies (VR/AR). The session featured keynote and pitch-style presentations, concluding with an interactive roundtable discussion to engage the audience.

Key Contributions

Keynote:
“Learning, upskilling and reskilling for new robotics and AI technologies”
– Presented by Alfio Minissale, COMAU
– The keynote highlighted the importance of workforce development in the era of AI-driven automation, showcasing training frameworks and innovative learning environments that support skill transformation and continuous upskilling in modern industrial settings.

Pitch Talks:

• From the Laboratory to the Plant: Integrating the Operator into Intelligent Automation in the Food Sector
  Ander Iriondo – TEKNIKER
  Presented human-centric approaches for integrating operators into intelligent automation workflows, demonstrating real-world applications and user acceptance in industrial environments.
• Show, Tell, and Deploy: Humanizing European Production through Imitation, VLMs, and Humanoids
  Urko Esnaola – TECNALIA
  Introduced novel methods for robot programming through demonstration, vision-language models, and humanoid systems, enabling more intuitive and human-oriented interaction with robotic platforms.
• Task Allocation Optimization in Hybrid Production Environments via Large Language Models
  Nikos Dimitropoulos – LMS, University of Patras
  Demonstrated the use of LLMs for optimizing task allocation between humans and robots, validated through real-world and simulated industrial scenarios.
• Empowering the Green Workforce: Hybrid Production Systems for the Future of Sustainable Re-X Processes
  Dionisis Andronas – LMS, University of Patras
  Addressed the challenges of circular manufacturing, emphasizing the role of human workers and AI-driven systems in enabling sustainable, flexible, and adaptive Re-X processes.
• VR/AR Assisted Robot Programming by Demonstration in Human-Robot Collaborative Scenarios
  Juan Francisco Blanes Noguera – Universitat Politècnica de València
  Presented immersive XR-based approaches for intuitive robot programming, enhancing human–robot collaboration through demonstration and digital twin environments.

Main Outcome

The workshop combined expert presentations with active audience engagement to explore key challenges and future directions in human-centric, AI-driven production systems. Live polling via Slido provided valuable insights from participants representing academia, industry, and research organizations, highlighting both technological opportunities and critical concerns around human–machine collaboration.

Questions and Results:

• What type of organization do you represent?
  • Academia – 43%
  • Industry (End User) – 0%
  • Industry (Technology Provider) – 22%
  • Research & Technology Organization (RTO) – 17%
  • Government – 13%
  • Other – 4%
• What is the key missing capability for true human–machine symbiosis in production?(Open responses)
  • “Mutual understanding”
  • “Trustworthy models of reality”
  • “Natural interaction”
  • “Safety of HRC workflows”
  • “Human acceptance”
  • “Lifelong reskilling”
• Where does ‘human-centricity’ currently fail in practice?
  • Used as a conceptual label rather than an engineering principle – 45%
  • Not a real priority in industrial deployment – 45%
  • Conflicts with productivity optimization – 30%
  • Reduced to ergonomics, ignoring cognition – 25%
  • Not measurable in system design – 25%
  • Other – 5%
• What is the real direction of current AI-driven production systems?
  • Humans are adapting to machine logic and constraints – 44%
  • Both trends are equally present – 25%
  • Machines are becoming more human-like – 13%
  • Neither – 13%
  • Other – 6%
• In practice, where should AI not be given decision authority? (Open responses)
  • “Life and death decisions”
  • “Final decision-making without human confirmation”
  • “Security and human resources”
  • “Creative tasks and innovation”
  • “Situations outside trained knowledge”
  • “Quality inspection and critical operations”

These results highlight a strong awareness within the community of the challenges in achieving true human–machine symbiosis. While there is clear progress in AI capabilities, participants emphasized the need for trust, transparency, and meaningful human involvement in system design. The findings also reflect a critical view of current “human-centric” approaches, pointing to a gap between conceptual frameworks and real industrial implementation, as well as ongoing concerns about the limits of AI autonomy in decision-making.