TY - JOUR
T1 - AI and the transformation of industrial work
T2 - Hybrid intelligence vs double-black box effect
AU - Wahlström, Mikael
AU - Tammentie, Bastian
AU - Salonen, Tuisku Tuuli
AU - Karvonen, Antero
N1 - Publisher Copyright:
© 2024
PY - 2024/7
Y1 - 2024/7
N2 - It is uncertain how the application of artificial intelligence (AI) technology transforms industrial work. We address this question from the perspective of cognitive systems, which, in this case, includes considerations of AI and process transparency, resilience, division of labor, and worker skills. We draw from a case study on glass tempering that includes a machine-vision-based quality control system and an advanced automation process control system. Based on task analysis and background literature, we develop the concept of hybrid intelligence that implies balanced AI transparency that supports upskilling and resilience. So-called fragmented intelligence, in turn, may result from the combination of the complexity of advanced automation along with the complexity of the process physics that places critical emphasis on expert knowledge. This combination can result in the so-called “double black box effect”, given that designing for understandability for the line workers might not be feasible: expert networks are needed for resilience.
AB - It is uncertain how the application of artificial intelligence (AI) technology transforms industrial work. We address this question from the perspective of cognitive systems, which, in this case, includes considerations of AI and process transparency, resilience, division of labor, and worker skills. We draw from a case study on glass tempering that includes a machine-vision-based quality control system and an advanced automation process control system. Based on task analysis and background literature, we develop the concept of hybrid intelligence that implies balanced AI transparency that supports upskilling and resilience. So-called fragmented intelligence, in turn, may result from the combination of the complexity of advanced automation along with the complexity of the process physics that places critical emphasis on expert knowledge. This combination can result in the so-called “double black box effect”, given that designing for understandability for the line workers might not be feasible: expert networks are needed for resilience.
KW - Artificial intelligence
KW - Glass tempering
KW - Hybrid intelligence
KW - Industrial processes
KW - Resilience
KW - Task analysis
UR - http://www.scopus.com/inward/record.url?scp=85189657398&partnerID=8YFLogxK
U2 - 10.1016/j.apergo.2024.104271
DO - 10.1016/j.apergo.2024.104271
M3 - Article
AN - SCOPUS:85189657398
SN - 0003-6870
VL - 118
JO - Applied Ergonomics
JF - Applied Ergonomics
M1 - 104271
ER -