In AI we trust? Perceptions about automated decision-making by artificial intelligence

Abstract

Fueled by ever-growing amounts of (digital) data and advances in artificial intelligence, decision-making in contemporary societies is increasingly delegated to automated processes. Drawing from social science theories and from the emerging body of research about algorithmic appreciation and algorithmic perceptions, the current study explores the extent to which personal characteristics can be linked to perceptions of automated decision-making by AI, and the boundary conditions of these perceptions, namely the extent to which such perceptions differ across media, (public) health, and judicial contexts. Data from a scenario-based survey experiment with a national sample (N = 958) show that people are by and large concerned about risks and have mixed opinions about fairness and usefulness of automated decision-making at a societal level, with general attitudes influenced by individual characteristics. Interestingly, decisions taken automatically by AI were often evaluated on par or even better than human experts for specific decisions. Theoretical and societal implications about these findings are discussed.

Artificial intelligence for political decision-making in the European Union: Effects on citizens’ perceptions of input, throughput, and output legitimacy

“…Some of the present results run counter to our hypotheses. Given the largely positive attitude to AI in the EU (European Commission, 2015, 2017), and in light of recent empirical evidence (Araujo et al, 2020;Marcinkowski et al, 2020), we expected ADM to score highly on output legitimacy. However, respondents expressed a more favorable view of HDM and HyDM outcomes, suggesting that they consider it illegitimate to leave important EU political decisions solely to automated systems.…”

“…Few studies have investigated the effects of ADM on perceptions of legitimacy, especially with respect to political decisions. To date, empirical studies have tended to focus on public sector areas, such as education and health, evaluating the effects of ADM as compared to human decision-making (HDM) in terms of variables, such as fairness and trust (Lee, 2018;Araujo et al, 2020;Marcinkowski et al, 2020). While those studies investigate decisions that affect individual citizens (e.g., decisions about loans or university admissions), the political context examined in this study refers to decisions that affect societal groups, or even society as a whole.…”

Towards Transparency by Design for Artificial Intelligence

“…Risk communication is traditionally an essential element of informed consent (Beauchamp and Childress 2001 ). Risks associated with automated decision-making have been discussed in the literature (Araujo et al 2020 ; Schermer 2011 ; Bahner et al 2008 ). Particularly, automated decision-making poses new risks of “privacy-invasive, discriminatory, and biased decision-making” (Wachter and Mittelstadt 2019 , p. 505), of reputation and informational self-determination, and the potential permanence and lack of contestability of problematic records.…”

The Individual and the Organizational Model of Quantum Decision-Making and Learning (MQDM&L): An Introduction and the application of the Quadruple Loop Learning

“…This results from an automation of decision-making (replacing humans) while increasing the rate of return on scale, scope, and investment in learning (Iansiti and Lakhani, 2020, p. 53). Notwithstanding, it can be assumed that more and more decisions will be taken by the AI algorithms and platforms and supported by big data (Araujo et al, 2020;Jarrahi, 2018), both horizontally and vertically (context specific, see example at Vaishya et al, 2020;and a review at Shahid, Rappon & Berta, 2019).…”