TY - GEN
T1 - Operational Criteria of Hybrid Intelligence for Generative AI Virtual Assistants
AU - Sherson, Jacob
AU - Rafner, Janet
AU - Büyükgüzel, Safinaz
N1 - Publisher Copyright:
© 2024 The Authors.
PY - 2024/6
Y1 - 2024/6
N2 - The concept of Hybrid Intelligence (HI) is frequently used interchangeably with Human-Centered AI (HCAI) and more broadly as human-in-the-loop. Dellerman et al. [1] outlined three differentiation criteria, emphasizing in particular the need for an evolving continuum of human-AI learning, a concept that has proven challenging to operationalize effectively. Recent efforts aim to expand the definition of HI beyond the domain of human-computer interaction to include application-oriented insights from management science [2]. This broader perspective integrates vital components such as facilitating end-user co-creation through narrative frameworks that foster psychological safety by addressing fears of job displacement [3,4], mitigating risks of deskilling during system deployment and scaling [5], and supporting business process innovation [2]. Additionally, in contrast to HCAI, the name hybrid intelligence conveys the possibly symmetric human-machine relationship and thereby preserves some of the disruptive potential of automated AI rather than relying on purely augmentation of human tasks and intentions [3]. Explicitly, the HI interaction should not only augment the existing, predefined task but also support aspects such as (business) process and business model re-engineering. Despite these considerations, a thorough discussion on which of the many established HCAI concepts and design guidelines form crucial components in achieving the aims of HI has so far been absent in literature. In particular, as it is becoming more and more likely that most knowledge workers will within a short timeframe become operators of complex virtual assistants tapping into LLMs and natural language interfaces, it becomes urgent to ensure that the human-ai interface and associated narrative is constructed to support HI principles and objectives. To initiate this discussion, we formulate explicitly updated HI design criteria in particular for generative AI virtual assistant design and discuss relevant HCAI concept.
AB - The concept of Hybrid Intelligence (HI) is frequently used interchangeably with Human-Centered AI (HCAI) and more broadly as human-in-the-loop. Dellerman et al. [1] outlined three differentiation criteria, emphasizing in particular the need for an evolving continuum of human-AI learning, a concept that has proven challenging to operationalize effectively. Recent efforts aim to expand the definition of HI beyond the domain of human-computer interaction to include application-oriented insights from management science [2]. This broader perspective integrates vital components such as facilitating end-user co-creation through narrative frameworks that foster psychological safety by addressing fears of job displacement [3,4], mitigating risks of deskilling during system deployment and scaling [5], and supporting business process innovation [2]. Additionally, in contrast to HCAI, the name hybrid intelligence conveys the possibly symmetric human-machine relationship and thereby preserves some of the disruptive potential of automated AI rather than relying on purely augmentation of human tasks and intentions [3]. Explicitly, the HI interaction should not only augment the existing, predefined task but also support aspects such as (business) process and business model re-engineering. Despite these considerations, a thorough discussion on which of the many established HCAI concepts and design guidelines form crucial components in achieving the aims of HI has so far been absent in literature. In particular, as it is becoming more and more likely that most knowledge workers will within a short timeframe become operators of complex virtual assistants tapping into LLMs and natural language interfaces, it becomes urgent to ensure that the human-ai interface and associated narrative is constructed to support HI principles and objectives. To initiate this discussion, we formulate explicitly updated HI design criteria in particular for generative AI virtual assistant design and discuss relevant HCAI concept.
UR - http://www.scopus.com/inward/record.url?scp=85198707927&partnerID=8YFLogxK
U2 - 10.3233/FAIA240229
DO - 10.3233/FAIA240229
M3 - Article in proceedings
AN - SCOPUS:85198707927
T3 - Frontiers in Artificial Intelligence and Applications
SP - 475
EP - 477
BT - HHAI 2024
A2 - Lorig, Fabian
A2 - Tucker, Jason
A2 - Lindstrom, Adam Dahlgren
A2 - Dignum, Frank
A2 - Murukannaiah, Pradeep
A2 - Theodorou, Andreas
A2 - Yolum, Pinar
PB - IOS Press BV
T2 - 3rd International Conference on Hybrid Human-Artificial Intelligence, HHAI 2024
Y2 - 10 June 2024 through 14 June 2024
ER -