No Algorithm Aversion in Improving AI
Uloženo v:
| Název: | No Algorithm Aversion in Improving AI |
|---|---|
| Autoři: | Wojciech Milczarski, Anna Borkowska, Emilia Biesiada, Laura Russak |
| Zdroj: | Collective and Individual Decisions. 37 |
| Informace o vydavateli: | Kozminski University, 2025. |
| Rok vydání: | 2025 |
| Popis: | Algorithm aversion is the tendency to avoid using algorithms or AI systems. Most of the stud-ies only present participants with the most recent performance of humans or AI. Through a series of experiments, involving N=905 participants, we investigated how people evaluate AI versus human performance, focusing on consistent high-quality output versus improvement over time. Our findings reveal that the preference for humans over AI significantly decreases when both demonstrate improvement. We observe this in the creative domain of tattoo design and other non-creative domains, such as law, logistics, and sales. Emphasizing AI’s improve-ment could effectively reduce algorithm aversion and increase AI acceptance. Our research contributes to understanding AI acceptance in domains, even those traditionally dominated by human creativity, offering insights for implementing AI systems in various fields. Algorithm aversion is the tendency to avoid using algorithms or AI systems. Most of the stud-ies only present participants with the most recent performance of humans or AI. Through a series of experiments, involving N=905 participants, we investigated how people evaluate AI versus human performance, focusing on consistent high-quality output versus improvement over time. Our findings reveal that the preference for humans over AI significantly decreases when both demonstrate improvement. We observe this in the creative domain of tattoo design and other non-creative domains, such as law, logistics, and sales. Emphasizing AI’s improve-ment could effectively reduce algorithm aversion and increase AI acceptance. Our research contributes to understanding AI acceptance in domains, even those traditionally dominated by human creativity, offering insights for implementing AI systems in various fields. |
| Druh dokumentu: | Article |
| ISSN: | 3071-7973 |
| DOI: | 10.7206/cid.3071-7973.9 |
| Přístupové číslo: | edsair.doi...........b1043f1d070b34efcdcf09d3b10b6fe9 |
| Databáze: | OpenAIRE |
Nájsť tento článok vo Web of Science | Abstrakt: | Algorithm aversion is the tendency to avoid using algorithms or AI systems. Most of the stud-ies only present participants with the most recent performance of humans or AI. Through a series of experiments, involving N=905 participants, we investigated how people evaluate AI versus human performance, focusing on consistent high-quality output versus improvement over time. Our findings reveal that the preference for humans over AI significantly decreases when both demonstrate improvement. We observe this in the creative domain of tattoo design and other non-creative domains, such as law, logistics, and sales. Emphasizing AI’s improve-ment could effectively reduce algorithm aversion and increase AI acceptance. Our research contributes to understanding AI acceptance in domains, even those traditionally dominated by human creativity, offering insights for implementing AI systems in various fields. Algorithm aversion is the tendency to avoid using algorithms or AI systems. Most of the stud-ies only present participants with the most recent performance of humans or AI. Through a series of experiments, involving N=905 participants, we investigated how people evaluate AI versus human performance, focusing on consistent high-quality output versus improvement over time. Our findings reveal that the preference for humans over AI significantly decreases when both demonstrate improvement. We observe this in the creative domain of tattoo design and other non-creative domains, such as law, logistics, and sales. Emphasizing AI’s improve-ment could effectively reduce algorithm aversion and increase AI acceptance. Our research contributes to understanding AI acceptance in domains, even those traditionally dominated by human creativity, offering insights for implementing AI systems in various fields. |
|---|---|
| ISSN: | 30717973 |
| DOI: | 10.7206/cid.3071-7973.9 |