To Err Is AI! Debugging as an Intervention to Facilitate Appropriate Reliance on AI Systems
A crowdsourced study (N=234) finds that an explanation-based debugging intervention failed to calibrate laypeople's estimates of AI performance and instead reduced reliance on a deceptive-review-detection AI.- doi
- 10.1145/3648188.3675130
- isbn
- 979-8-4007-0595-3
- name
- To Err Is AI! Debugging as an Intervention to Facilitate Appropriate Reliance on AI Systems
- source
- publisher_html+pdf_figure_crops
- acm_url
- https://dl.acm.org/doi/10.1145/3648188.3675130
- authors
- Gaole He, Abri Bharos, Ujwal Gadiraju
- doi_url
- https://doi.org/10.1145/3648188.3675130
- license
- CC BY 4.0
- summary
- A crowdsourced study (N=234) finds that an explanation-based debugging intervention failed to calibrate laypeople's estimates of AI performance and instead reduced reliance on a deceptive-review-detection AI.
- arxiv_url
- https://arxiv.org/abs/2409.14377
- published
- 2024-09-10
- conference
- HT '24: 35th ACM Conference on Hypertext and Social Media, Poznan, Poland, September 10-13, 2024
- open_access
- true
- acm_html_url
- https://dl.acm.org/doi/full/10.1145/3648188.3675130
- displayAuthor
- Gaole He, Abri Bharos, and Ujwal Gadiraju (Delft University of Technology, Delft, The Netherlands)
- displayPublishTime
- 2024-09-10
Powered by Seed HypermediaOpen App