Table 2 Perspectives for managing... | Emerald Publishing

Table 2

Perspectives for managing explainability of AI

Perspectives	Description	Key sources
Pragmatism in explainability	Understanding explainability relationally and assessing the trustworthiness of AI in its communication with humans based on contextual and other factors that mediate the value of the explainability of the communication	Felzmann et al. (2019), Garibaldi (2019), Miller (2019), Robbins (2019b)
Contextualization of the explanation	Assessing tasks, capabilities and expectations of the AI system based on its context, which is everything that shapes and influences our perceptions, cognition and actions in a particular domain	Bellotti and Edwards (2001), European Commission (2019), Lawless et al. (2019), Lecue (2019), Miller (2019)
Cohabitation of human agency and AI agency	Examining how responsibility for AI decisions and consequences should be distributed between humans and the AI system	Adadi and Berrada (2018), Coeckelbergh (2009, 2020), D'Acquisto (2020), Kroll (2018)
Metrics and standardization	Context-specific as well as universal measures that can compare, evaluate and quantify explainability methods and their efficiency, outcomes and impacts	Arrieta et al. (2020), Doshi-Velez and Kim (2017), Garibaldi (2019), Gunning and Aha (2019), Gunning et al. (2019), Preece (2018)
Regulatory and ethical principles	Ethical and legal frameworks that ensure and promote human safety and autonomy in interactions with AI and maintain and encourage visibility and explainability in the use and propagation of AI systems	Bertino et al. (2019), Hacker et al. (2020), Robbins (2019b), Stahl and Wright (2018), Tóth (2019)
Other emerging solutions	Innovative and emerging solutions, particularly incorporating AI enveloping, blockchain and fuzzy systems, for improving and managing explainability in AI	Bertino et al. (2019), Fernandez et al. (2019), Floridi (2011), Garibaldi (2019), Kshetri (2019), Nassar et al. (2020), Robbins (2019b)

Perspectives	Description	Key sources
Pragmatism in explainability	Understanding explainability relationally and assessing the trustworthiness of AI in its communication with humans based on contextual and other factors that mediate the value of the explainability of the communication	Felzmann et al. (2019), Garibaldi (2019), Miller (2019), Robbins (2019b)
Contextualization of the explanation	Assessing tasks, capabilities and expectations of the AI system based on its context, which is everything that shapes and influences our perceptions, cognition and actions in a particular domain	Bellotti and Edwards (2001), European Commission (2019), Lawless et al. (2019), Lecue (2019), Miller (2019)
Cohabitation of human agency and AI agency	Examining how responsibility for AI decisions and consequences should be distributed between humans and the AI system	Adadi and Berrada (2018), Coeckelbergh (2009, 2020), D'Acquisto (2020), Kroll (2018)
Metrics and standardization	Context-specific as well as universal measures that can compare, evaluate and quantify explainability methods and their efficiency, outcomes and impacts	Arrieta et al. (2020), Doshi-Velez and Kim (2017), Garibaldi (2019), Gunning and Aha (2019), Gunning et al. (2019), Preece (2018)
Regulatory and ethical principles	Ethical and legal frameworks that ensure and promote human safety and autonomy in interactions with AI and maintain and encourage visibility and explainability in the use and propagation of AI systems	Bertino et al. (2019), Hacker et al. (2020), Robbins (2019b), Stahl and Wright (2018), Tóth (2019)
Other emerging solutions	Innovative and emerging solutions, particularly incorporating AI enveloping, blockchain and fuzzy systems, for improving and managing explainability in AI	Bertino et al. (2019), Fernandez et al. (2019), Floridi (2011), Garibaldi (2019), Kshetri (2019), Nassar et al. (2020), Robbins (2019b)