A Vice Framework for AI

TY - GEN

T1 - A Vice Framework for AI

AU - Chubb, Jennifer Alison

AU - Forstenzer, Josh

AU - Kidd, Ian

N1 - Philosophy of AI Conference (PhAI 2023) to be published in forthcming journal.

PY - 2023/11/15

Y1 - 2023/11/15

N2 - In this work, we consider if we myopically focus on the moral positives of AI, then there is a risk of inadvertently creating corrupting techno-social systems, which is potentially going to make our future work environments even worse than they’re already likely to be. Contemporary theorising about the ethical dimensions of AI indicates at least two problematic tendencies. The first, a default focus on concepts like autonomy, justice, fairness, non-discrimination, explicable in terms of the entrenchment of a liberal morality of the sort central to much contemporary moral and political philosophy. Second, a focus on concepts that track the positive dimensions of moral practice and experience, such as virtue, flourishing, moral progress, the good life (Vallor, 2016). Taken together, the result is large-scale moral frameworks for moral appraisal of new technologies that fail to systematically address the negative aspects of moral life (Jobin et al, 2019): one is emphasis on virtues and excellences to the occlusion of vices and failings; another is affirmation of the edifying potential of new technologies to the relative neglect of the corrupting possibilities (Coekelberg, 2020). p.59). We suggest that a more comprehensive assessment of the moral possibilities of new technologies in the future of work requires a more complex set of concepts and sensibilities and that there are good empirical and theoretical reasons to consider the vicious and corrupting dimensions of new technologies. If so, exploring the negative moral possibilities of these technologies requires a greater sensitivity to what we will call technomoral-vices.

AB - In this work, we consider if we myopically focus on the moral positives of AI, then there is a risk of inadvertently creating corrupting techno-social systems, which is potentially going to make our future work environments even worse than they’re already likely to be. Contemporary theorising about the ethical dimensions of AI indicates at least two problematic tendencies. The first, a default focus on concepts like autonomy, justice, fairness, non-discrimination, explicable in terms of the entrenchment of a liberal morality of the sort central to much contemporary moral and political philosophy. Second, a focus on concepts that track the positive dimensions of moral practice and experience, such as virtue, flourishing, moral progress, the good life (Vallor, 2016). Taken together, the result is large-scale moral frameworks for moral appraisal of new technologies that fail to systematically address the negative aspects of moral life (Jobin et al, 2019): one is emphasis on virtues and excellences to the occlusion of vices and failings; another is affirmation of the edifying potential of new technologies to the relative neglect of the corrupting possibilities (Coekelberg, 2020). p.59). We suggest that a more comprehensive assessment of the moral possibilities of new technologies in the future of work requires a more complex set of concepts and sensibilities and that there are good empirical and theoretical reasons to consider the vicious and corrupting dimensions of new technologies. If so, exploring the negative moral possibilities of these technologies requires a greater sensitivity to what we will call technomoral-vices.

KW - AI

KW - ETHICS

KW - VICE

KW - VIRTUE

KW - FRAMEWORK

M3 - Conference contribution

BT - Philosophy of AI Conference (PhAI 2023)

T2 - Philosophy of AI Conference (PhAI 2023)

Y2 - 15 December 2023 through 17 December 2023

ER -