Who needs explanation and when? Juggling explainable AI and user epistemic uncertainty

Jiang, Jinglu; Kahai, Surinder S.; Yang, Ming

doi:10.1016/j.ijhcs.2022.102839

Cited by 37 publications

(9 citation statements)

References 77 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The results of the interviews indicated that the need for explanations was mainly in cases where AI outcomes were new or anomalous to respondents. This finding is also in line with previous studies' findings outside AR [67,102].…”

Section: Resultssupporting

confidence: 94%

“…(3) Reliability. Ensuring the reliability of AI outcomes is essential for non-trivial decision-making processes so that users can rely on a trustworthy system [102,124,178], e.g., daily activity recommendations for personal health management or automatic emergency service contacting in safety-threatening incidents. (4) Informativeness.…”

Section: Platform-agnostic Factorsmentioning

confidence: 99%

“…Earlier XAI research aims to help AI/ML developers on model debugging (e.g., [106,140,178,179,240]) or assist domain experts such as clinicians by revealing more information such as causality and certainty (e.g., [75,129,217,222]). Recently, there has been a growing amount of XAI research focusing on the non-expert end-users [31,74,102]. Existing studies have found that XAI can help end-users resolve confusion and build trust [67,170].…”

Section: Introductionmentioning

confidence: 99%

“…After selecting the appropriate content, default explanations need to be concise and can be further simplified by highlighting the most important types [26,40]. General end-users are primarily interested in Why, which is in line with experts' advice (Suggestion 3 about default options) and previous literature [102,132,133]. Designers' can leverage their expertise to determine whether other types should be omitted or combined with Why in a specific context.…”

mentioning

confidence: 97%

See 3 more Smart Citations

XAIR: A Framework of Explainable AI in Augmented Reality

Yu²,

Jonker

et al. 2023

Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems

View full text Add to dashboard Cite

Explainable AI (XAI) has established itself as an important component of AI-driven interactive systems. With Augmented Reality (AR) becoming more integrated in daily lives, the role of XAI also becomes essential in AR because end-users will frequently interact with intelligent services. However, it is unclear how to design effective XAI experiences for AR. We propose XAIR, a design framework that addresses when, what, and how to provide explanations of AI output in AR. The framework was based on a multi-disciplinary literature review of XAI and HCI research, a large-scale survey probing 500+ end-users' preferences for AR-based explanations, and three workshops with 12 experts collecting their insights about XAI design in AR. XAIR's utility and effectiveness was verified via a study with 10 designers and another study with 12 end-users. XAIR can provide guidelines for designers, inspiring them to identify new design opportunities and achieve effective XAI designs in AR.

show abstract

Section: Resultssupporting

confidence: 94%

Section: Platform-agnostic Factorsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

mentioning

confidence: 97%

See 2 more Smart Citations

XAIR: A Framework of Explainable AI in Augmented Reality

Yu²,

Jonker

et al. 2023

Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems

View full text Add to dashboard Cite

show abstract

“…To delineate the association between uncertainty and trust it is important to discern two distinct types of uncertainty: epistemic uncertainty and ontic uncertainty. Epistemic uncertainty is characterized by an observer’s lack of knowledge about the state of a system at a particular time, which can be resolved through acquiring additional information or knowledge (Busemeyer et al, 2020; Jiang et al, 2022). Ontic uncertainty describes a person’s internal uncertainty regarding a specific response, such as a decision amongst different options and may only be resolved through some sort of interaction(s) (Busemeyer et al, 2020; Busemeyer & Bruza, 2014; Kvam et al, 2021).…”

Section: Trust In Human-ai Decision-makingmentioning

confidence: 99%

Temporal Evolution of Trust in Artificial Intelligence-Supported Decision-Making

Humr,

Canan,

Demir

2023

Proceedings of the Human Factors and Ergonomics Society Annual

View full text Add to dashboard Cite

AI is set to take over some tasks within the decision space that have traditionally been reserved for humans. In return, human decision-makers interacting with AI systems entails rationalization of AI outputs by humans, who may have difficulty forming trust around such AI-generated information. Although a variety of analytical methods have provided some insights into human trust in AI, a more comprehensive understanding of trust may be augmented by generative theories that capture the temporal evolution of trust. Therefore, an open system modeling approach, representing trust as a function of time with a single probability distribution, can potentially improve modeling human trust in an AI system. Results of this study could improve machine behaviors that may help steer a human’s preference to a more Bayesian optimal rationality which is useful in stressful decision-making scenarios.

show abstract

Reconciling the personalization–privacy paradox via DoctorBots: The roles of service robot acceptance model elements and technology anxiety

Shi,

Lu,

Zhou

2023

J of Consumer Behaviour

View full text Add to dashboard Cite

The role of artificial intelligence (AI)‐based DoctorBots in improving healthcare and the medical industry is expected to increase significantly in the coming decades. However, contrary to the general view that AI can gradually replace human work, user acceptance of DoctorBots has become an obstacle to the development of AI medical diagnosis. Building on the service robot acceptance model (sRAM), this study investigates the potential of the functional, socioemotional, and relational elements of DoctorBots to reconcile the personalization–privacy paradox, thus enhancing user acceptance. Via two scenario‐based experiments with 398 participants, this study reveals that the negative influence of the personalization–privacy paradox on user acceptance is exacerbated when users' technology anxiety is high. In addition, an online survey of 400 DoctorBot users indicates that ease of use, subjective social norms, social presence, and rapport are effective in addressing both nonpersonalization (NPC) and privacy concerns (PVC). These findings suggest that the healthcare industry can leverage DoctorBots to implement self‐diagnosis. Specifically, DoctorBots' functional elements are effective in mitigating users' NPC, and their relational elements are effective in extenuating users' PVC.

show abstract

Who needs explanation and when? Juggling explainable AI and user epistemic uncertainty

Cited by 37 publications

References 77 publications

XAIR: A Framework of Explainable AI in Augmented Reality

XAIR: A Framework of Explainable AI in Augmented Reality

Temporal Evolution of Trust in Artificial Intelligence-Supported Decision-Making

Reconciling the personalization–privacy paradox via DoctorBots: The roles of service robot acceptance model elements and technology anxiety

Contact Info

Product

Resources

About