The 5Cs—opposing effects of AI explainability
| Opposing effects | Description | Key sources |
|---|---|---|
| Comprehensibility | Understandability of what explainability of AI means to different stakeholders, especially laypeople | Adadi and Berrada (2018), Arrieta et al. (2020), De Graaf and Malle (2017), Harper (2019), Holzinger et al. (2019), Miller (2019), Ntoutsi et al. (2020), Páez (2019) |
| Conduct | The innovativeness and performance of AI systems and their explainability | D'Acquisto (2020), Diez-Olivan et al. (2019), Gunning and Aha (2019), Gunning et al. (2019), Robbins (2019a), Silver et al. (2016) |
| Confidentiality | Confidentiality, security and safety risks due to the explainability of AI systems | Arrieta et al. (2020), Holzinger et al. (2019), Tóth (2019) |
| Completeness | Consequences due to AI logic that cannot be proved or explained | Arrieta et al. (2020), D'Acquisto (2020), Silver et al. (2016) |
| Confidence in AI | Overconfidence in AI outcomes vs. mistrust of AI due to its explanation | Bertino et al. (2019), Harper (2019), Mittelstadt et al. (2016), Pieters (2011) |
| Opposing effects | Description | Key sources |
|---|---|---|
| Comprehensibility | Understandability of what explainability of AI means to different stakeholders, especially laypeople | |
| Conduct | The innovativeness and performance of AI systems and their explainability | |
| Confidentiality | Confidentiality, security and safety risks due to the explainability of AI systems | |
| Completeness | Consequences due to AI logic that cannot be proved or explained | |
| Confidence in AI | Overconfidence in AI outcomes vs. mistrust of AI due to its explanation |
Sharing content requires targeting cookies to be enabled. Please update your cookie preferences to use this feature.