CITATION — REFERENCE ENTRY
To Trust or to Think: Cognitive Forcing Functions Can Reduce Overreliance on AI in AI-Assisted Decision-Making — Proceedings of the ACM on Human-Computer Interaction
- Key
- bucinca-et-al-2021-cognitive-forcing
- Authors
- Buçinca, Zana; Malaya, Maja Barbara; Gajos, Krzysztof Z.
- Issued
- 2021-4
- Type
- article-journal
- Container
- Proceedings of the ACM on Human-Computer Interaction
- Volume
- 5
- Issue
- CSCW1
- Pages
- 1-21
- DOI
- 10.1145/3449287
Raw CSL JSON
{
"DOI": "10.1145/3449287",
"URL": "https://www.eecs.harvard.edu/~kgajos/papers/2021/bucinca21trust.pdf",
"page": "1-21",
"type": "article-journal",
"issue": "CSCW1",
"title": "To Trust or to Think: Cognitive Forcing Functions Can Reduce Overreliance on AI in AI-Assisted Decision-Making",
"author": [
{
"given": "Zana",
"family": "Buçinca"
},
{
"given": "Maja Barbara",
"family": "Malaya"
},
{
"given": "Krzysztof Z.",
"family": "Gajos"
}
],
"issued": {
"date-parts": [
[
2021,
4
]
]
},
"number": "Article 188",
"volume": "5",
"container-title": "Proceedings of the ACM on Human-Computer Interaction"
}
Claims
-
Three cognitive forcing functions — requiring an initial decision before showing the AI suggestion, delaying the AI suggestion, or requiring the user to explicitly request it — significantly reduced overreliance on incorrect AI suggestions compared with plain explainable-AI conditions; participants rated the forcing-function conditions as more effortful and less preferred.
"Through two experiments (N=199 and N=302), we demonstrated that cognitive forcing interventions reduce overreliance on AI compared to the no-AI baseline and simple explainable AI approaches."
-
Cognitive forcing functions disproportionately benefited participants with high Need for Cognition, suggesting that interventions designed to reduce overreliance on AI may help most those who are already disposed to think analytically.
"Our results show that cognitive forcing functions disproportionately benefited participants with high Need for Cognition"
Available in