Socio-cognitive biases in folk AI ethics and risk discourse

Abstract

The ongoing conversation on AI ethics and politics is in full swing and has spread to the general public. Rather than contributing by engaging with the issues and views discussed, we want to step back and comment on the widening conversation itself. We consider evolved human cognitive tendencies and biases, and how they frame and hinder the conversation on AI ethics. Primarily, we describe our innate human capacities known as folk theories and how we apply them to phenomena of different implicit categories. Through examples and empirical findings, we show that such tendencies specifically affect the key issues discussed in AI ethics. The central claim is that much of our mostly opaque intuitive thinking has not evolved to match the nature of AI, and this causes problems in democratizing AI ethics and politics. Developing awareness of how our intuitive thinking affects our more explicit views will add to the quality of the conversation.

Responsible (use of) AI

“…In such situations, there may be no “good” option, but rather an actor must select between multiple options none of which are “morally-flawless” (Misselhorn, 2022 ). It is important to use the term “select” over terms such as “think” to avoid anthropomorphizing of AI—because regardless of one's views regarding the feasibility of creating a “conscious” machine in the long-term (see Dignum, 2017 ; Laakasuo et al, 2021 ), it is hard to challenge the fact that machines will be placed in situations where they must ingest data and act. Human understanding and acceptance of ethical behavior will shape how well AI is adopted in society, “…for the wider public to accept the proliferation of artificial intelligence-driven vehicles on their roads, both groups will need to understand the origins of the ethical principles that are programmed into these vehicles” (Awad et al, 2018 , p. 64).…”

Expectations of how machines use individuating information and base-rates

“…People are also reluctant to recommend computers as medical practitioners for others when those individuals are described as having unique medical conditions. Similarly, people may view machines as comparatively inflexible -they might believe that computers are capable of doing only what they have been programmed to do (Laakasuo et al, 2021; also see Kim & Duhachek 2020).…”