Similarity and Consistency in Algorithm-Guided Exploration
Yongping Bao,
Ludwig Danwitz,
Fabian Dvorak,
Sebastian Fehrler,
Lars Hornuf,
Hsuan Yu Lin and
Bettina von Helversen
No 10188, CESifo Working Paper Series from CESifo
Abstract:
Algorithmic advice holds the potential to significantly enhance human decision-making, particularly in dynamic and complex tasks involving a trade-off between exploration and exploitation. We investigate the conditions under which people are willing to accept advice from algorithms in such tasks, focusing on the interplay between individuals’ and the advising algorithm’s exploration preferences. In an online experiment, we engineered reinforcement learning algorithms to favor either exploration or exploitation and observed participants’ decision-making processes, modeling them using a cognitive framework comparable to the algorithm. Interestingly, individuals showed a greater inclination to follow the advice of exploitative, consistent algorithms, possibly perceiving algorithmic consistency as a sign of competence. They did not exhibit a preference for algorithms with similar exploration tendencies to their own. Moreover, we found that participants’ exploration tendencies influenced the behavior of the algorithms, underscoring the importance of considering the mutually reinforcing co-behaviors of algorithms and humans.
Keywords: algorithms; decision support systems; recommender systems; advice-taking; multi-armed bandit; search; exploration-exploitation; cognitive modeling (search for similar items in EconPapers)
JEL-codes: C91 D83 (search for similar items in EconPapers)
Date: 2022
New Economics Papers: this item is included in nep-cbe, nep-cmp, nep-dcm and nep-exp
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (1)
Downloads: (external link)
https://www.cesifo.org/DocDL/cesifo1_wp10188.pdf (application/pdf)
Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.
Export reference: BibTeX
RIS (EndNote, ProCite, RefMan)
HTML/Text
Persistent link: https://EconPapers.repec.org/RePEc:ces:ceswps:_10188
Access Statistics for this paper
More papers in CESifo Working Paper Series from CESifo Contact information at EDIRC.
Bibliographic data for series maintained by Klaus Wohlrabe ().