AI & Society最新文献

The term ‘artificial intelligence’ has arguably come to function in political discourse as, what Laclau called, an ‘empty signifier’. This article traces the shifting political discourse on AI within three key institutions of global governance–OHCHR, WHO, and UNESCO–and, in so doing, highlights the role of ‘crisis’ moments in justifying a series of pivotal re-articulations. Most important has been the attachment of AI to the narrative around digital automation in human healthcare. Greatly enabled by the societal context of the pandemic, all three institutions have moved from being critical of the unequal power relations in the economy of AI to, today, reframing themselves primarily as facilitators tasked with helping to ensure the application of AI technologies. The analysis identifies a shift in which human health and healthcare is framed as in a ‘crisis’ to which AI technology is presented as the remedy. The article argues the need to trace these discursive shifts as a means by which to understand, monitor, and where necessary also hold to account these changes in the governance of AI in society.

Algorithmic decision-making systems have the potential to amplify existing discriminatory patterns and negatively affect perceptions of justice in society. There is a need for a revision of mechanisms to address discrimination in light of the unique challenges presented by these systems, which are not easily auditable or explainable. Research efforts to bring fairness to ADM solutions should be viewed as a matter of justice and trust among actors should be ensured through technology design. Ideas that move us to explore the notions of justice within the field of political thinking, aiming to identify the elements that frame the social context of the discriminatory decisions produced by algorithmic decision-making solutions. Our explorations suggest that the efforts to bring fairness to ADMs should be seen as a matter of justice and that trust in these systems can be ensured through careful technology design.

Automation is inherently tied to ethical challenges because of its potential involvement in morally loaded decisions. In the present research, participants (n = 34) took part in a moral multi-trial dilemma-based task where the agent (human vs. machine) and the behavior (action vs. inaction) factors were randomized. Self-report measures, in terms of morality, consciousness, responsibility, intentionality, and emotional impact evaluation were gathered, together with electroencephalography (delta, theta, beta, upper and lower alpha, and gamma powers) and peripheral autonomic (electrodermal activity, heart rate variability) data. Data showed that moral schemata vary as a function of the involved decider, and when the agent and behavior factors are crossed. Subjects did not consider machines full moral deciders to the same degree as humans and tend to morally better accept human active behavior and machine inaction. Moreover, the autonomic physiological activity might support the a-posteriori moral evaluation. In the evaluation of the agent’s consciousness, a beta ventrolateral prefrontal synchronization was detected for human action and machine inaction, while a generalized gamma synchronization occurred in artificial agent trials while rating the emotional impact of the decider’s behavior. The detected differences might point to a potential explicit and implicit asymmetry in moral reasoning toward artificial and human agents.

In informationally mature societies, almost all organisations record, generate, process, use, share and disseminate data. In particular, the rise of AI and autonomous systems has corresponded to an improvement in computational power and in solving complex problems. However, the resulting possibilities have been coupled with an upsurge of ethical risks. To avoid the misuse, underuse, and harmful use of data and data-based systems like AI, we should use an ethical framework appropriate to the object of its reasoning. Unfortunately, in recent years, the space for data-related ethics has not been precisely defined in organisations. As a consequence, there has been an overlapping of responsibilities and a void of clear accountabilities. Ethical issues have, therefore, been dealt with using inadequate levels of abstraction (e.g. legal, technical). Yet, if building an ethical infrastructure requires the collaboration of each body, addressing ethical issues related to data requires leaving room for the appropriate level of abstraction. This paper first aims to show how the space of data ethics is already latent in organisations. It then highlights how to redefine roles (chief data ethics officer, data ethics committee, etc.) and codes (code of data ethics) to create and maintain an environment where ethical reasoning about data, information, and AI systems may flourish.

The humanities and social sciences (HSS) are being turned upside down by advances in artificial intelligence (AI), and their very existence could be threatened. These sciences are being profoundly destabilised by a dual process of naturalisation of social phenomena and fetishisation of numbers, accentuated by the development of AI (part 1). Both STM (science, technology, medicine) and HSS are facing major epistemological challenges, but for the latter they carry the risk of marginalisation (part 2). The humanities and social sciences remain the best equipped to question the social construct represented by the development of AI. However, this essential approach is not enough. We need to ask ourselves: how can the HSS reintroduce interpretation when they have less and less control over how data is put together? Only a balanced partnership between STM and HSS is likely to meet all these challenges (part 3). Using the case of education, which has long been at the forefront of developments in other sectors of social life, we would like to show how and on what priority issues such a partnership can be built (part 4).