Bad machines corrupt good morals

Abstract

Machines powered by Artificial Intelligence (AI) are now influencing the behavior of humans in ways that are both like and unlike the ways humans influence each other. In light of recent research showing that other humans can exert a strong corrupting influence on people's ethical behavior, worry emerges about the corrupting power of AI agents. To estimate the empirical validity of these fears, we review the available evidence from behavioral science, human-computer interaction, and AI research. We propose that the main social roles through which both humans and machines can influence ethical behavior are (a) role model, (b) advisor, (c) partner, and (d) delegate. When AI agents become influencers (role models or advisors), their corrupting power may not exceed (yet) the corrupting power of humans. However, AI agents acting as enablers of unethical behavior (partners or delegates) have many characteristics that may let people reap unethical benefits while feeling good about themselves, indicating good reasons for worry. Based on these insights, we outline a research agenda that aims at providing more behavioral insights for better AI oversight.

Fooled twice: People cannot detect deepfakes but think they can

“…As detecting deepfakes appears less a matter of motivation and attention, our results suggest that deepfakes warrant special attention in the research and policy field of digital misinformation. In line with the widely voiced concern about deepfakes becoming a new AI threat ( Caldwell et al., 2020 ; Chesney and Citron, 2019 ; Köbis et al., 2021 ), our findings suggest some of the previously established strategies against manipulation do not hold for detection of deepfakes. A need exists for more research on human-centered strategies against deepfakes.…”

“…Together, these findings suggest that people apply an overly optimistic seeing-is-believing heuristic, which might put them at a particular risk of being influenced by deepfakes. We discuss these findings below, emphasizing the need for more research on the interplay of behavior by humans and machines ( Rahwan et al., 2019 ; Köbis et al., 2021 ).…”

“…Applications like Face swapping GAN are democratizing the creation of deepfakes ( Nirkin et al., 2019 ). Identifying who is behind the production of a particular deepfake becomes increasingly demanding, effectively undermining accountability and often resulting in impunity ( King et al., 2020 ; Köbis et al., 2021 ). At the same time, deepfakes can reach an unprecedented audience ( Diakopoulos and Johnson, 2019 ).…”

“…Since false information often spreads faster and permeates social media networks deeper than accurate information ( Vosoughi et al., 2018 ), correcting falsehoods like deepfakes online presents a complex challenge ( Korshunov and Marcel, 2018 ). Therefore, worry emerges about using deepfakes as a new deceptive tool ( Köbis et al., 2021 ), for example, in mass voter misinformation campaigns before elections ( Dobber et al., 2021 ).…”

The Conflict Between People’s Urge to Punish AI and Legal Systems

“… 1 We use the term “machine” as a interchangeable term for AI systems and robots, i.e., embodied forms of AI. Recent work on the human factors of AI systems has used this term to refer to both AI and robots (e.g., ( Köbis et al, 2021 )), and some of the literature that has inspired this research uses similar terms when discussing both entities, e.g., ( Matthias, 2004 ). …”