Research on the ethics of algorithms has grown substantially over the past decade. Alongside the exponential development and application of machine learning algorithms, new ethical problems and solutions relating to their ubiquitous use in society have been proposed. This article builds on a review of the ethics of algorithms published in 2016, 2016). The goals are to contribute to the debate on the identification and analysis of the ethical implications of algorithms, to provide an updated analysis of epistemic and normative (...) concerns, and to offer actionable guidance for the governance of the design, development and deployment of algorithms. (shrink)

Research on the ethics of algorithms has grown substantially over the past decade. Alongside the exponential development and application of machine learning algorithms, new ethical problems and solutions relating to their ubiquitous use in society have been proposed. This article builds on a review of the ethics of algorithms published in 2016, 2016). The goals are to contribute to the debate on the identification and analysis of the ethical implications of algorithms, to provide an updated analysis of epistemic and normative (...) concerns, and to offer actionable guidance for the governance of the design, development and deployment of algorithms. (shrink)

In this commentary, we discuss the nature of reversible and irreversible questions, that is, questions that may enable one to identify the nature of the source of their answers. We then introduce GPT-3, a third-generation, autoregressive language model that uses deep learning to produce human-like texts, and use the previous distinction to analyse it. We expand the analysis to present three tests based on mathematical, semantic, and ethical questions and show that GPT-3 is not designed to pass any of them. (...) This is a reminder that GPT-3 does not do what it is not supposed to do, and that any interpretation of GPT-3 as the beginning of the emergence of a general form of artificial intelligence is merely uninformed science fiction. We conclude by outlining some of the significant consequences of the industrialisation of automatic and cheap production of good, semantic artefacts. (shrink)

Deep learning is currently the most prominent and widely successful method in artificial intelligence. Despite having played an active role in earlier artificial intelligence and neural network research, philosophers have been largely silent on this technology so far. This is remarkable, given that deep learning neural networks have blown past predicted upper limits on artificial intelligence performance—recognizing complex objects in natural photographs and defeating world champions in strategy games as complex as Go and chess—yet there remains no universally accepted explanation (...) as to why they work so well. This article provides an introduction to these networks as well as an opinionated guidebook on the philosophical significance of their structure and achievements. It argues that deep learning neural networks differ importantly in their structure and mathematical properties from the shallower neural networks that were the subject of so much philosophical reflection in the 1980s and 1990s. The article then explores several different explanations for their success and ends by proposing three areas of inquiry that would benefit from future engagement by philosophers of mind and science. (shrink)

Advances in artificial intelligence and robotics stand to change many aspects of our lives, including our values. If trends continue as expected, many industries will undergo automation in the near future, calling into question whether we can still value the sense of identity and security our occupations once provided us with. Likewise, the advent of social robots driven by AI, appears to be shifting the meaning of numerous, long-standing values associated with interpersonal relationships, like friendship. Furthermore, powerful actors’ and institutions’ (...) increasing reliance on AI to make decisions that may affect how people live their lives may have a significant impact on privacy while also raising issues about algorithmic transparency and human control. In this paper, building and expanding on previous works, we will look at how the deployment of Artificial Intelligence technology may lead to changes in identity, security, and other crucial values (such as friendship, fairness, and privacy). We will discuss what challenges we may face in the process, while critically reflecting on whether such changes may be desirable. Finally, drawing on a series of considerations underlying virtue ethics, we will formulate a set of preliminary suggestions, which—we hope—can be used to more carefully guide the future roll out of AI technologies for human flourishing; that is, for social and moral good. (shrink)

An AI winter may be defined as the stage when technology, business, and the media come to terms with what AI can or cannot really do as a technology without exaggeration. Through discussion of previous AI winters, this paper examines the hype cycle (which by turn characterises AI as a social panacea or a nightmare of apocalyptic proportions) and argues that AI should be treated as a normal technology, neither as a miracle nor as a plague, but rather as of (...) the many solutions that human ingenuity has managed to devise. (shrink)

An increasing number of technology firms are implementing processes to identify and evaluate the ethical risks of their systems and products. A key part of these review processes is to foresee potential impacts of these technologies on different groups of users. In this article, we use the expression Ethical Foresight Analysis to refer to a variety of analytical strategies for anticipating or predicting the ethical issues that new technological artefacts, services, and applications may raise. This article examines several existing EFA (...) methodologies currently in use. It identifies the purposes of ethical foresight, the kinds of methods that current methodologies employ, and the strengths and weaknesses of each of these current approaches. The conclusion is that a new kind of foresight analysis on the ethics of emerging technologies is both feasible and urgently needed. (shrink)

This article contributes to the debate around the abilities of large language models such as GPT-3, dealing with: firstly, evaluating how well GPT does in the Turing Test, secondly the limits of such models, especially their tendency to generate falsehoods, and thirdly the social consequences of the problems these models have with truth-telling. We start by formalising the recently proposed notion of reversible questions, which Floridi & Chiriatti propose allow one to ‘identify the nature of the source of their answers’, (...) as a probabilistic measure based on Item Response Theory from psychometrics. Following a critical assessment of the methodology which led previous scholars to dismiss GPT’s abilities, we argue against claims that GPT-3 completely lacks semantic ability. Using ideas of compression, priming, distributional semantics and semantic webs we offer our own theory of the limits of large language models like GPT-3, and argue that GPT can competently engage in various semantic tasks. The real reason GPT’s answers seem senseless being that truth-telling is not amongst them. We claim that these kinds of models cannot be forced into producing only true continuation, but rather to maximise their objective function they strategize to be plausible instead of truthful. This, we moreover claim, can hijack our intuitive capacity to evaluate the accuracy of its outputs. Finally, we show how this analysis predicts that a widespread adoption of language generators as tools for writing could result in permanent pollution of our informational ecosystem with massive amounts of very plausible but often untrue texts. (shrink)

In this article, we shed light on the emergence, diffusion, and use of socio-technological future visions. The artificial intelligence future vision of the German federal government is examined and juxtaposed with the respective news media coverage of the German media. By means of a content analysis of frames, it is demonstrated how the German government strategically uses its AI future vision to uphold the status quo. The German media largely adapt the government´s frames and do not integrate alternative future narratives (...) into the public debate. These findings are substantiated in the framing of AI futures in policy documents of the German government and articles of four different German newspapers. It is shown how the German past is mirrored in the German AI future envisioned by the government, safeguarding the present power constellation that is marked by a close unity of politics and industry. The German media partly expose the government´s frames and call for future visions that include fundamentally different political designs less influenced by the power structures of the past and present. (shrink)

It has been suggested that to overcome the challenges facing the UK’s National Health Service (NHS) of an ageing population and reduced available funding, the NHS should be transformed into a more informationally mature and heterogeneous organisation, reliant on data-based and algorithmically-driven interactions between human, artificial, and hybrid (semi-artificial) agents. This transformation process would offer significant benefit to patients, clinicians, and the overall system, but it would also rely on a fundamental transformation of the healthcare system in a way that (...) poses significant governance challenges. In this article, we argue that a fruitful way to overcome these challenges is by adopting a pro-ethical approach to design that analyses the system as a whole, keeps society-in-the-loop throughout the process, and distributes responsibility evenly across all nodes in the system. (shrink)

Is ChatGPT a good teacher? Or could it be? As understanding is widely acknowledged as one of the fundamental aims of education, the answer to these questions depends on whether ChatGPT fosters or could foster the acquisition of understanding in its users. In this paper, I tackle this issue in two steps. In the first part of the paper, I explore and analyze the set of skills and social-epistemic virtues that a teacher must exemplify to perform her job well – (...) in those contexts in which epistemic aims are at play and in which understanding plays a pivotal role. In the second part of the paper, I put my conception of good teacher to test and deal with the question whether, and to which extent, the software ChatGPT is (or could serve as, if suitably modified or fine-tuned) a good teacher for its users. I close with some final reflections that point to further directions of research. -/- . (shrink)

Intelligibility and interpretability related to artificial intelligence (AI) are crucial for enabling explicability, which is vital for establishing constructive communication and agreement among various stakeholders, including users and designers of AI. It is essential to overcome the challenges of sharing an understanding of the details of the various structures of diverse AI systems, to facilitate effective communication and collaboration. In this paper, we propose four fundamental terms: “I/O,” “Constraints,” “Objectives,” and “Architecture.” These terms help mitigate the challenges associated with intelligibility (...) and interpretability in AI by providing appropriate levels of abstraction to describe structure of AI systems generally, thereby facilitating the sharing of understanding among various stakeholders. The relationship between the Objective of AI designers and the Purpose of AI users is linked to the issues of AI alignment. (shrink)

Embodied cognition is an interpretative—or hermeneutical—cognition inherent in motor-sensory perception intrinsically informed by biological and sociocultural memory, a cognition embedded in the organism as well as the socio-cultural environment interacting with it (Ward et al. TOPOI 36:365–375, 2017), of which technologies are a part. Yet, smart machines are advancing on human abilities to perceive and interpret concerning the accuracy, quantity, and quality of the data processed. Machines process and categorize images, perform classification tasks, they calculate and perform pattern analysis, all (...) machine learning processes are task-specific. Machine learning processes resemble human interpretative processes; however, these are two very different ways of “dealing with something,” although both can be said to be hermeneutical, one enactive, the other material; the first is a meaning-generating interpretative process, the second a statistics-based information output. The information extracted from all the data fed into the computer is the end-product of machine learning, whereas human interpretative enactments are continuous and necessary for any application of the information output. (shrink)

A variety of approaches have appeared in academic literature and in design practice representing “ethics-first” methods. These approaches typically focus on clarifying the normative dimensions of design, or outlining strategies for explicitly incorporating values into design. While this body of literature has developed considerably over the last 20 years, two themes central to the endeavour of ethics and values in design (E + VID) have yet to be systematically discussed in relation to each other: (a) designer agency, and (b) the (...) strength of normative claims informing the design process. To address this gap, we undertook a structured review of leading E + VID approaches and critiques, and classified them according to their positions on normative strength, and views regarding designer agency. We identified 18 distinct approaches and 13 critiques that met the inclusion criteria for our review. Included papers were distributed across the spectrum of views regarding normative strength, and we found that no approaches and only one critique represented a view characteristic of “low” designer agency. We suggest that the absence of “low” designer agency approaches results in the neglect of crucial influences on design as targets of intervention by designers. We conclude with suggestions for future research that might illuminate strategies to achieve ethical design in information mature societies, and argue that without attending to the tensions raised by balancing normatively “strong” visions of the future with limitations imposed on designer agency in corporate-driven design settings, “meaningful” ethical design will continue to encounter challenges in practice. (shrink)

This paper considers a host of definitions and labels attached to the concept of smart cities to identify four dimensions that ground a review of ethical concerns emerging from the current debate. These are: network infrastructure, with the corresponding concerns of control, surveillance, and data privacy and ownership; post-political governance, embodied in the tensions between public and private decision-making and cities as post-political entities; social inclusion, expressed in the aspects of citizen participation and inclusion, and inequality and discrimination; and sustainability, (...) with a specific focus on the environment as an element to protect but also as a strategic element for the future. Given the persisting disagreements around the definition of a smart city, the article identifies in these four dimensions a more stable reference framework within which ethical concerns can be clustered and discussed. Identifying these dimensions makes possible a review of the ethical implications of smart cities that is transversal to their different types and resilient towards the unsettled debate over their definition. (shrink)

Since the publication of ChatGPT and Dall-E, there has been heavy discussions on the possible dangers of generative artificial intelligence (AI) for society. These discussions question the extent to which the development of AI can be regulated by politics, law, and civic actors. An important arena for discourse on AI is the news media. The news media discursively construct AI as a technology that is more or less possible to regulate. There are various reasons for an assumed regulatability. Some voices (...) highlight the predominant capitalism of modern society as an ideology that enforces an uncontrolled development of AI. Others call on the rationality of civic society, which will push AI toward a development that serves humankind. Summarized, the discourse on AI floats between positions of technological and social determinism. The article conducts a discourse analysis on 113 articles from the German newspapers Süddeutsche Zeitung and Frankfurter Allgemeine Zeitung. The analysis shows how these center-left and center-right media frame the AI development in terms of social and technological determinism. As these newspapers reach out to a large audience, the article shows the kinds of perspectives on AI that confront civic society in Germany every day. News media can motivate or impede social action, as they frame the potential necessity of society and its members to intervene against certain developments. The article shows how the newspapers promote an understanding of AI, by which citizens will feel motivated to insist on a regulation of AI by politics and law. (shrink)

Digital technologies create and shape our environments, the infosphere, where we spend increasingly more time. Through exploration of such concepts as "latency", "real time" and "unreal time", this article discusses how time has changed due to the digital revolution over the past half-century.

Practical wisdom enables moral decision-making and action by aligning one’s apprehension of proximate goods with a distal, socially embedded interpretation of a more ultimate Good. A focus on purpose within the overall process mutually informs human moral psychology and moral AI development in their examinations of practical wisdom. AI practical wisdom could ground an AI system’s apprehension of reality in a sociotechnical moral process committed to orienting AI development and action in light of a pluralistic, diverse interpretation of that Good. (...) Defining AI purpose in terms of the human struggle to determine the moral “ought” (versus the way reality is) can overcome practical and normative limitations of principle-based approaches to AI ethics without requiring idiosyncratic moral reasoning or a premature commitment to a single ethical theory, which could hinder navigating the moral implications of future unknowns. Dividing practical wisdom into proximate, mediating, and distal processes enables each division to incorporate the moral dimensions of the situation, individual AI system, and sociotechnical community, respectively. Drawing upon externalist (4E) and ecological cognitive psychology theories to extend Xavier Zubiri’s sentient intelligence, proximate apprehension of what an object affords the perceiver includes affordances for practical and moral purposes. Those apprehensions are structured by moral and other social-cognitive schemas that mediate moral constructs, the apprehended situation, and individual commitments and other motivating identifications in a process Ignacio Ellacuria calls historical reality. Distally, Josiah Royce’s community of interpretation characterizes one way of organizing a sociotechnical process to synthesize diverse interpretations of the Good, orient AI moral practice, and align AI development with human and AI moral purpose. (shrink)

In September 2020, Helsinki and Amsterdam announced the launch of their open AI registers—the first cities in the world to offer such a service. The AI registers describe what, where, and how AI applications are being used in the two municipalities; how algorithms were assessed for potential bias or risks; and how humans use the AI services. Examining issues of security and transparency, this paper discusses the potential for implementing AI in an urban public service setting and how this might (...) be harnessed to improve services and citizens’ experiences. -/- . (shrink)

Nowadays, new directions for the development of artificial intelligence (AI) have emerged, the task has been set to develop artificial general intelligence (AGI), which is able to go beyond the narrow AI, gain a high degree of autonomy, independently solve problems in different environmental conditions and thus have the ability to perform the functions of natural intelligence. In this regard, important philosophical, theoretical, and methodological questions arise concerning the definition and evaluation of the social significance of new AI achievements, especially (...) regarding the correlation of their socio-humanitarian and technological aspects. It is necessary to overcome a purely technocratic approach, which usually ignores the negative consequences of digitalization, the possible risks and threats of AI development. These difficulties are conditioned by the paradigm gap between the system of concepts used to describe technologies and the systems of concepts specific to socio-humanitarian descriptions and explanations. The so-called ontological paradox arises during digital transformation and the introduction of AI into social systems. An ontological paradox is also an epistemological paradox, since any statement of an ontological type presupposes its epistemological justification. To overcome this gap, V.E. Lepskiy proposes the conception of self-developing polysubjective environments – cybernetics of the third order. This conception allows to create a conceptual “bridge” between two systems of definitions that do not have direct logical connections between them. A productive tool for this can be the information approach, which is widely used to solve such interdisciplinary problems. This makes possible to theoretically correctly connect technological and socio-humanitarian descriptions in a single conceptual structure and to analyze main socio-humanitarian criteria for evaluating digital innovations. The article discusses in detail the main provisions of the information approach and its application for the development of AGI systems, the socio-humanitarian assessment of AI autonomous development, and the resolution of security issues. (shrink)

Our social relations are changing, we are now not just talking to each other, but we are now also talking to artificial intelligence (AI) assistants. We claim AI assistants present a new form of digital connectivity risk and a key aspect of this risk phenomenon is related to user risk awareness (or lack of) regarding AI assistant functionality. AI assistants present a significant societal risk phenomenon amplified by the global scale of the products and the increasing use in healthcare, education, (...) business, and service industry. However, there appears to be little research concerning the need to not only understand the changing risks of AI assistant technologies but also how to frame and communicate the risks to users. How can users assess the risks without fully understanding the complexity of the technology? This is a challenging and unwelcome scenario. AI assistant technologies consist of a complex ecosystem and demand explicit and precise communication in terms of communicating and contextualising the new digital risk phenomenon. The paper then argues for the need to examine how to best to explain and support both domestic and commercial user risk awareness regarding AI assistants. To this end, we propose the method of creating a risk narrative which is focused on temporal points of changing societal connectivity and contextualised in terms of risk. We claim the connectivity risk narrative provides an effective medium in capturing, communicating, and contextualising the risks of AI assistants in a medium that can support explainability as a risk mitigation mechanism. (shrink)

Artificial Intelligence (AI) in the 21st century is no longer perceived as a purely technological phenomenon, more and more becoming a social and humanitarian phenomenon that develops in a complex context of cultural, value, philosophical, and ethical aspects of human life. The impact of AIrelated technologies on contemporary society is still difficult to assess fully, which does not prevent enthusiastic researchers and political leaders from attempting to define a value framework that will ensure the use of AI for societal development. (...) As interest in AI grows, more and more technologically advanced countries in the world are creating their own strategies for the development and use of this technological marvel of the 21st century. These pioneering documents often seem vague and indefinite, but nevertheless they allow us to assess how the political and scientific and technological elites of these countries see the value orientations of AI technology development, both nationally and internationally. The article presents a philosophical analysis of the value orientations of AI technology development in the USA and China – modern scientific and technological leaders in this field – on the basis of strategic documents defining the development and application of AI in these countries from the position of post-non-classical scientific rationality. The authors of the article conclude that, at the contemporary post-non-classical stage of science development, the value component is not only one of the integral components of scientific and technological activities but may be decisive in determining the goals and objectives of high-tech development at the state level. (shrink)