ATB organises special session on AI-driven Sliding Work Sharing at PRO-VE 2025

At the PRO-VE 2025 Conference on Hybrid Human-AI Collaborative Networks, held in Porto, Portugal, from 27 to 29 October 2025, ATB organised the special session “AI-driven Sliding Work Sharing for Human-Robot / Human-AI Collaboration.” The session brought together research addressing how humans, artificial intelligence and robots can collaborate more effectively in dynamic working environments.

The special session was organised by Ana Correia and Sebastian Scholze from ATB in connection with the European research projects AI4Work and SOPRANO and was chaired by Sebastian Scholze. Its focus closely reflects the challenges addressed by both projects: developing human-centred approaches in which AI and robotic systems support people while the level of automation and human involvement can adapt to the situation.

A central topic of the session was Sliding Work Sharing (SWS). Rather than assigning a task permanently either to a human or to an automated system, SWS allows the balance between human and machine involvement to change during operation. Relevant factors can include the current operational context, the confidence or uncertainty of the AI system, and the experience and availability of the human operator. This concept is at the core of the AI4Work approach to adaptive human-AI and human-robot collaboration.

Context-aware Sliding Work Sharing for logistics

During the session, ATB presented the paper “Context Aware Sliding Work Sharing for Human-AI Collaboration in Logistics Domain”, authored by Sebastian Scholze, Ana Correia and Gunnar Große Hovest.

The paper proposes a framework that combines context awareness with Sliding Work Sharing to dynamically determine an appropriate level of human and AI involvement. The approach uses information about the current situation together with factors such as machine confidence and human expertise to support decisions about whether an AI system should act autonomously, keep a human informed, request human confirmation or leave the decision entirely to the human.

The approach is illustrated through a logistics scenario, where operational conditions can change quickly and AI-supported decisions therefore need to take the current context into account. Context modelling, monitoring and extraction are used to transform operational data into information that can support adaptive work-sharing decisions.

This reflects a broader objective of AI4Work: rather than maximising automation as an end in itself, the project investigates how work can be shared dynamically between people, AI systems and robots according to the needs of the current situation. In the logistics pilot, for example, contextual information from yard-management processes can be used to support scheduling and other operational decisions while retaining an appropriate degree of human oversight.

The PRO-VE session provided an opportunity to discuss these approaches with researchers working on human-AI collaboration, human-robot interaction and adaptive automation, and to connect the work carried out in AI4Work and SOPRANO with the wider research community.

The paper is published in the PRO-VE 2025 proceedings by Springer:

Sebastian Scholze, Ana Correia and Gunnar Große Hovest: “Context Aware Sliding Work Sharing for Human-AI Collaboration in Logistics Domain.”
DOI: 10.1007/978-3-032-05681-8_24