Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Factory Operators' Perspectives on Cognitive Assistants for Knowledge Sharing: Challenges, Risks, and Impact on Work

Samuel Kernan Freire, Tianhao He, Chaofan Wang, Evangelos Niforatos, Alessandro Bozzon

TL;DR

This paper investigates how cognitive assistants (CAs), including smartphone voice interfaces and LLM-powered chatbots, affect knowledge sharing in manufacturing through a two-year longitudinal, phase-based field study in two European detergent factories. Using a hybrid deductive/inductive thematic analysis of 251 operator/manager comments, the study finds that CAs can speed issue resolution and enable knowledge transfer, especially for novices, but raise concerns about privacy, surveillance, data quality, and integration with existing workflows. It develops design guidelines for rapid information retrieval, scalable knowledge updates, and human-centered governance to address socio-technical tensions. The work contributes to CSCW and knowledge-management literature by detailing real-world adoption dynamics, operator-management tensions, and practical considerations for aligning infrastructure, data practices, and user education with AI-assisted knowledge sharing on the shop floor.

Abstract

In the shift towards human-centered manufacturing, our two-year longitudinal study investigates the real-world impact of deploying Cognitive Assistants (CAs) in factories. The CAs were designed to facilitate knowledge sharing among factory operators. Our investigation focused on smartphone-based voice assistants and LLM-powered chatbots, examining their usability and utility in a real-world factory setting. Based on the qualitative feedback we collected during the deployments of CAs at the factories, we conducted a thematic analysis to investigate the perceptions, challenges, and overall impact on workflow and knowledge sharing. Our results indicate that while CAs have the potential to significantly improve efficiency through knowledge sharing and quicker resolution of production issues, they also introduce concerns around workplace surveillance, the types of knowledge that can be shared, and shortcomings compared to human-to-human knowledge sharing. Additionally, our findings stress the importance of addressing privacy, knowledge contribution burdens, and tensions between factory operators and their managers.

Factory Operators' Perspectives on Cognitive Assistants for Knowledge Sharing: Challenges, Risks, and Impact on Work

TL;DR

This paper investigates how cognitive assistants (CAs), including smartphone voice interfaces and LLM-powered chatbots, affect knowledge sharing in manufacturing through a two-year longitudinal, phase-based field study in two European detergent factories. Using a hybrid deductive/inductive thematic analysis of 251 operator/manager comments, the study finds that CAs can speed issue resolution and enable knowledge transfer, especially for novices, but raise concerns about privacy, surveillance, data quality, and integration with existing workflows. It develops design guidelines for rapid information retrieval, scalable knowledge updates, and human-centered governance to address socio-technical tensions. The work contributes to CSCW and knowledge-management literature by detailing real-world adoption dynamics, operator-management tensions, and practical considerations for aligning infrastructure, data practices, and user education with AI-assisted knowledge sharing on the shop floor.

Abstract

In the shift towards human-centered manufacturing, our two-year longitudinal study investigates the real-world impact of deploying Cognitive Assistants (CAs) in factories. The CAs were designed to facilitate knowledge sharing among factory operators. Our investigation focused on smartphone-based voice assistants and LLM-powered chatbots, examining their usability and utility in a real-world factory setting. Based on the qualitative feedback we collected during the deployments of CAs at the factories, we conducted a thematic analysis to investigate the perceptions, challenges, and overall impact on workflow and knowledge sharing. Our results indicate that while CAs have the potential to significantly improve efficiency through knowledge sharing and quicker resolution of production issues, they also introduce concerns around workplace surveillance, the types of knowledge that can be shared, and shortcomings compared to human-to-human knowledge sharing. Additionally, our findings stress the importance of addressing privacy, knowledge contribution burdens, and tensions between factory operators and their managers.
Paper Structure (53 sections, 4 figures, 2 tables)

This paper contains 53 sections, 4 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Background and Related Work
  3. Knowledge Management
  4. Knowledge Sharing Systems
  5. Integrating AI in Knowledge Sharing Systems
  6. Cognitive Assistants for Factory Operators
  7. AI Assistants
  8. Knowledge-intensive AI Assistants
  9. Cognitive Assistants: Acquiring and Sharing Human Knowledge
  10. Socio-technical Challenges
  11. Real-World Case Studies in Factories
  12. Approach and Context
  13. Systems, Protocols, and Participants
  14. Phase One: Technology Probe during Production Line Operations
  15. Phase Two: CA with a Simulated Production Line
  16. ...and 38 more sections

Figures (4)

  • Figure 1: Headset used at the production line for data collection during phase 1
  • Figure 2: (a) The Rasa X user interface from phase 2 and (b) A simulated "weight checker" user interface that was connected to the assistant during phase 2. Each bar represents the weight of a canister that was filled with detergent. The "weight checker" weighs each canister and stops the production line if the weight falls below a specified threshold.
  • Figure 3: (a) Android user interface for phase 3 and 4 and (b) The query tab of the LLM-powered CA from phase 4
  • Figure 4: A sticker used to block the stereoscopic camera for anonymous human tracking that overlooks the production line.