A “psychopathic” Artificial Intelligence: the possible risks of a deviating AI in Education

Abstract

This work analyses the use of artificial intelligence in education from an interdisciplinary point of view. New studies demonstrated that an AI can “deviate” and become potentially malicious, due to programmers’ biases, corrupted feeds or purposeful actions. Knowing the pervasive use of artificial intelligence systems, including in the educational environment, it seemed necessary to investigate when and how an AI in education could deviate. We started with an investigation of AI and the risks it poses, wondering if they could be applied also to educative AI. We then reviewed the increasing literature that deals with the use of technology in the classroom, and the criticism about it, referring to specific use cases. Finally, as a result, the authors formulate questions and suggestions for further research, to bridge conceptual gaps underlined by lack of research.

The Dark Path to Eternal Life: Machiavellianism Predicts Approval of Mind Upload Technology

“…Contrary to our preregistered prediction, attitudes toward mind upload were best predicted by Machiavellianism and not by psychopathy. However, given the conceptual and empirical links between these components of the Dark Triad, our results nevertheless corroborate the concern echoed in AI safety scholarship: namely, that mind upload may selectively appeal to antisocial individuals with morally suspect aims (Sotala & Yampolskiy, 2015; and others as well: Zanetti et al, 2019;Althaus & Bauman, 2020;MacAskill, 2020;Bostrom, 2014).…”

“…Regarding the failure of our pre-registered hypothesis that psychopathy specifically would be associated with mind upload approval and utilitarianism, we would like to note that Machiavellianism and psychopathy are substantially correlated and their items can load on a mutual factor (Jones & Paulhus, 2014). We focused on psychopathy, since it was the emphasis of recent scholarship A(G)I safety (Sotala & Yampolskiy 2015; see also Zanetti et al, 2019;Althaus & Bauman, 2020;MacAskill, 2020). Upon deeper reflection, since psychopathy in the Dark Triad model is associated with impulsivity and immediate need for gratification, this trait is likely not associated with liking a technology that could bring one power over others in the long term.…”

“…The existential risk of AGIs (through mind upload or otherwise) is that a psychopathic or otherwise anti-social individual might aspire to be the first person to create an AGI or to be the first one to be "uploaded" and develop superhuman cognitive abilities (Bostrom, 2014;Sotala & Gloor, 2017;Sotala & Yampolskiy, 2015;Yampolskiy, 2018). Naturally, if transhumanists are interested in mind upload technology because of its potential to alleviate suffering and help us possibly cope with the problems of Artificial Superintelligence, it is indeed a problem if these technologies fall into the hands of those individuals, who would use them to subjugate others or to increase the total amount of suffering in other ways (Sotala & Gloor, 2017;Sotala & Yampolskiy, 2015;Zanetti et al, 2019;Althaus & Bauman, 2020;MacAskill, 2020).…”

Power to the Teachers: An Exploratory Review on Artificial Intelligence in Education

“…The epitomised hierarchy and taxonomy of AI and how it impacts technological and social practices may be further delimited using the Substitution, Augmentation, Modification Redefinition (SAMR) model, developed by Puentedura [13] as a developmental ontological framework that demonstrates how AI will increasingly influence the dynamics of technology development as a means of entering a state of transformational activity underpinned by advanced forms of human intelligence (see Figure 4). The 'interdisciplinary' nature of AI in terms of emulating how the human mind processes information and knowledge from a cognitive and socio-cultural perspective has been embraced by Zanetti et al [14] and Dodigovic [15] referring to AI as an interdisciplinary area of knowledge and research, whose aim is to understand how the human mind works and then emulate this understanding to AI technology design. Dodigovic argues that a fundamental factor for AI to accomplish such emulations is the knowledge of language.…”

Influence of Artificial Intelligence in Education on Adolescents’ Social Adaptability: A Machine Learning Study

“…In addition, Holstein [10] pointed out that the need of teachers and students toward AIEd remains unclear [10,11]. Zanetti, Iseppi, and Cassese [12] also raised their apprehension that AIEd may be "deviate" and become potentially malicious, due to programmers' biases or other purposeful actions. Therefore, it is worth discussing the potential influence of AIEd.…”