There has been considerable optimistic speculation on how well ChatGPT-4 would perform in a Turing Test. However, no minimally serious implementation of the test has been reported to have been carried out. This brief note documents the re-sults of subjecting ChatGPT-4 to 10 Turing Tests, with different interrogators and participants. The outcome is tremendously disappointing for the optimists. Despite ChatGPT reportedly outperforming 99.9% of humans in a Verbal IQ test, it falls short of passing the Turing Test. In 9 out (...) of the 10 tests conducted, the interroga-tors successfully identified ChatGPT-4 and the human participant. The probability of obtaining this result from a process in which the interrogator is really no better than chance at correct identification is calculated to be less than 1%. An additional question was posed to the interrogators at the end of each test: What led them to distinguish between the human and the machine? The interrogators, who effectively filtered out ChatGPT-4 from passing the Turing Test for intelligence, stated that they could identify the machine because it, in effect, responded more intelligently than the human. Subsequently, ChatGPT-4 was tasked with differentiating syntax from semantics and self-corrected when falling for the fallacy of equivocation. The curious situation is arrived at that passing the Turing Test for intelligence remains a challenge that ChatGPT-4 has yet to overcome, precisely because, as per the inter-rogators, its intellectual abilities surpass those of individual humans. (shrink)

This paper provides a philosophically informed and robust account of the global justice implications of Artificial Intelligence (AI). We first discuss some of the key theories of global justice, before justifying our focus on the Capabilities Approach as a useful framework for understanding the context-specific impacts of AI on lowto middle-income countries. We then highlight some of the harms and burdens facing low- to middle-income countries within the context of both AI use and the AI supply chain, by analyzing the (...) extraction of materials, which includes mineral extraction and the environmental harms associated with it, and the extraction of labor, which includes unethical labor practices, low wages, and the trauma experienced by some AI workers. We then outline some of the potential harms and benefits that AI poses, how these are distributed, and what global justice implications this has for low- to middle-income countries. Finally, we articulate the global justice significance of AI by utilizing the Capabilities Approach. We argue that AI must be considered from a global justice perspective given that, globally, AI puts significant downward pressure on several elements of well-being thereby making it harder for people to achieve threshold levels of the central human capabilities needed for a life of dignity. (shrink)

This article traces the historical development of the ethics of emerging technologies. It argues that during the late 2000s and 2010s, the field of ethics of technology transformed from a fragmented, reactive, and methodologically underdeveloped discipline focused on mature technologies and lacking policy orientation into a more cohesive, proactive, methodologically sophisticated, and policy-focused field with a strong emphasis on emerging technologies. An agenda for this transition was set in Jim Moor’s seminal publication “Why We Need Better Ethics for Emerging Technologies”.

‘Are There Decisions Computers Should Never Make?’ is one of James H. Moor’s many groundbreaking papers in computer ethics, and it is one that I have thought a good deal about since its publication in 1979 and especially in recent years in relation to current discourse on AI. In this paper, I describe Jim’s analysis, reflect on its relevance to current thinking about AI, and take issue with several of his arguments. The conclusion of Jim’s paper is that computers should (...) never choose human values and goals. I suggest that this is not possible because of the nature of values and how they are intertwined in computer decision making. (shrink)

This article challenges the dominant ‘black box’ metaphor in critical algorithm studies by proposing a phenomenological framework for understanding how social media algorithms manifest themselves in user experience. While the black box paradigm treats algorithms as opaque, self-contained entities that exist only ‘behind the scenes’, this article argues that algorithms are better understood as genetic phenomena that unfold temporally through user-platform interactions. Recent scholarship in critical algorithm studies has already identified various ways in which algorithms manifest in user experience: through (...) affective responses, algorithmic self-reflexivity, disruptions of normal experience, points of contention, and folk theories. Yet, while these studies gesture toward a phenomenological understanding of algorithms, they do so without explicitly drawing on phenomenological theory. This article demonstrates how phenomenology, particularly a Husserlian genetic approach, can further conceptualize these already-documented algorithmic encounters. Moving beyond both the paradigm of artifacts and static phenomenological approaches, the analysis shows how algorithms emerge as inherently relational processes that co-constitute user experience over time. By reconceptualizing algorithms as genetic phenomena rather than black boxes, this paper provides a theoretical framework for understanding how algorithmic awareness develops from pre-reflective affective encounters to explicit folk theories, while remaining inextricably linked to users’ self-understanding. This phenomenological framework contributes to a more nuanced understanding of algorithmic mediation in contemporary social media environments and opens new pathways for investigating digital technologies. (shrink)

The advent of quantum computing will compromise current asymmetric cryptography. Awaiting this moment, global superpowers are routinely collecting and storing encrypted data, so as to later decrypt it once sufficiently strong quantum computers are in place. We argue that this situation gives rise to a new mode of global surveillance that we refer to as a _quantum panopticon._ Unlike traditional forms of panoptic surveillance, the quantum panopticon introduces a _temporal axis_, whereby data subjects’ future pasts can be monitored from an (...) unknown “superposition” in the quantum future_._ It also introduces a new level of _uncertainty,_ in that the future watchman’s very existence becomes a function of data subjects’ efforts to protect themselves from being monitored in the present. Encryption may work as a momentary protection, but increases the likelihood of long-term preservation for future decryption, because encrypted data is stored longer than plaintext data. To illustrate the political and ethical aspects of these features, we draw on cryptographic as well as theoretical surveillance literature and call for urgent consideration of the wider implications of quantum computing for the global surveillance landscape. (shrink)

An individual human has value partly in virtue of their uniqueness. Personal avatar technology—technology which creates a digital replication of a real person—appears to have the potential to undermine that value. Here I explore if and how avatars might make humans less valuable by undermining the value that a human gains from being unique. Ultimately, I conclude that, while avatars cannot make humans no longer unique, they could significantly undermine the value that we place on human uniqueness. First, I argue (...) that a qualitative model of uniqueness cannot account for the unique value that a person has. This leads to the significant and surprising claim that necessarily unique properties of humans cannot accommodate the value arising from human uniqueness: humans have unique value in virtue of being contingently irreplaceable. I explore how the use of personal avatars might undermine or even destroy that value. Finally, I consider further applications of the theory of unique human value, including how it might explain and accommodate our attachment to personal avatars themselves. (shrink)

This paper is intended as a tribute to the late James Moor. An esteemed Dartmouth professor, who published in many areas of philosophy, including logic, Moor is perhaps best remembered today for his pioneering work in the field of computer ethics. His seminal (and award-winning) article, “What Is Computer Ethics?” (_Metaphilosophy_, 1985) was highly influential both in defining and shaping the then nascent field of computer ethics. Many other computer-ethics-related papers followed over the next quarter century, in which Moor examined (...) a range of topics – from moral responsibility to autonomy to privacy in the context of computing and emerging technologies, including nanotechnology and AI. And while the insights and frameworks put forth in many of his published works have received the acclaim they deserve, Moor’s contribution to the privacy literature remains, in my view, underappreciated. In trying to show why his privacy theory deserves much more attention than received to date, I also briefly describe the evolution of Moor’s position on privacy – from his earlier publications on that topic to a comprehensive and systematic privacy framework. I then suggest that a further exploration of his privacy theory would benefit researchers working in technology ethics in general, and AI ethics in particular. Finally, I encourage privacy scholars to take a closer look at Moor’s privacy framework to see whether they might be able to tease out and disclose some potential insights and features that may still be embedded in that robust theory of privacy. (shrink)

Around the turn of this century a number of emerging technologies were in the news, raising some potentially significant ethical questions. Given that they were emerging they as yet had no, or very few, impacts, so it was not obvious how best to assess them ethically. Jim Moor addressed this issue and offered three suggestions for a better ethics for emerging technologies. His first was that ethics should be dynamic, that is, it should be an ongoing process before, during and (...) after the technological development. Second, there should be close collaboration between the researchers and developers on the one hand, and ethicists and social scientists on the other. Finally, ethical analyses should be more sophisticated. In this paper I argue that environmental issues and the questioning of core ethical values should be a central part of the ethics of emerging technologies, using AI examples. Given the kind of beings that we are, technology and the environment are closely connected for human flourishing. (shrink)

This article delivers an account of what it is for a physical system to be programmable. Despite its significance in computing and beyond, today’s philosophical discourse on programmability is impoverished. This contribution offers a novel definition of _physical programmability_ as the degree to which the selected operations of an automaton can be reconfigured in a controlled way. The framework highlights several key insights: the constrained applicability of physical programmability to material automata, the characterization of selected operations within the neo-mechanistic framework, (...) the understanding of controlled reconfiguration through the causal theory of interventionism, and the recognition of physical programmability as a gradual notion. The account can be used to individuate programmable (computing) systems and taxonomize concrete systems based on their programmability. The article closes by posing some open questions and offering avenues for future research in this domain. (shrink)

We study a controversial application of algorithmic profiling in the public sector, the Austrian AMAS system. AMAS was supposed to help caseworkers at the Public Employment Service (PES) Austria to allocate support measures to job seekers based on their predicted chance of (re-)integration into the labor market. Shortly after its release, AMAS was criticized for its apparent unequal treatment of job seekers based on gender and citizenship. We systematically investigate the AMAS model using a novel real-world dataset of young job (...) seekers from Vienna, which allows us to provide the first empirical evaluation of the AMAS model with a focus on fairness measures. We further apply bias mitigation strategies to study their effectiveness in our real-world setting. Our findings indicate that the prediction performance of the AMAS model is insufficient for use in practice, as more than 30% of job seekers would be misclassified in our use case. Further, our results confirm that the original model is biased with respect to gender as it tends to (incorrectly) assign women to the group with high chances of re-employment, which is not prioritized in the PES’ allocation of support measures. However, most bias mitigation strategies were able to improve fairness without compromising performance and thus may form an important building block in revising profiling schemes in the present context. (shrink)

There are more than 350,000 health apps available in public app stores. The extolled benefits of health apps are numerous and well documented. However, there are also concerns that poor-quality apps, marketed directly to consumers, threaten the tenets of evidence-based medicine and expose individuals to the risk of harm. This study addresses this issue by assessing the overall quality of evidence publicly available to support the effectiveness claims of health apps marketed directly to consumers. To assess the quality of evidence (...) available to the public to support the effectiveness claims of health apps marketed directly to consumers, an audit was conducted of a purposive sample of apps available on the Apple App Store. We find the quality of evidence available to support the effectiveness claims of health apps marketed directly to consumers to be poor. Less than half of the 220 apps (44%) we audited state that they have evidence to support their claims of effectiveness and, of these allegedly evidence-based apps, more than 70% rely publicly on either very low or low-quality evidence. For the minority of app developers that do publish studies, significant methodological limitations are commonplace. Finally, there is a pronounced tendency for apps—particularly mental health and diagnostic apps—to either borrow evidence generated in other (typically offline) contexts or to rely exclusively on unsubstantiated, unpublished user metrics as evidence to support their effectiveness claims. Health apps represent a significant opportunity for individual consumers and healthcare systems. Nevertheless, this opportunity will be missed if the health apps market continues to be flooded by poor quality, poorly evidenced, and potentially unsafe apps. It must be accepted that a continuing lag in generating high-quality publicly available evidence of app effectiveness and safety is not inevitable: it is a choice. Just because it will be challenging to raise the quality of the evidence base publicly available to support the claims of health apps, this does not mean that the bar for evidence quality should be lowered. Innovation for innovation’s sake must not be prioritized over public health and safety. (shrink)

The paper examines twelve legal regimes related to the governance and regulation of both the environmental risks and opportunities brought forth by the use of AI systems and AI models in the Artificial Intelligence Act (‘AIA’) of EU law. The assessment of risks and opportunities of AI related to the environment includes the high-risk management procedures under Art. 9 of the AIA, the “fundamental rights impact assessment” of Art. 27, and the codes of conduct of Art. 95. These provisions are (...) supplemented by further regulatory regimes, such as the proposal of EU directive on sustainable consumption and green claims, and Reg. (EU) 2023/588 on environmental and space sustainability, among others. The aim of the analysis is to specify which are the less or the more environmentally friendly regulatory regimes set up with the AIA. The claim is that Art. 9, 27 and 95 are among the less green pieces of the whole legislation. (shrink)

Most people who have considered the matter have concluded that machines cannot be moral agents. Responsibility for acting on the outputs of machines must always rest with a human being. A key problem for the ethical use of AI, then, is to ensure that it does not block the attribution of responsibility to humans or lead to individuals being unfairly held responsible for things over which they had no control. This is the “responsibility gap”. In this paper, we argue that (...) the claim that machines cannot be held responsible for their actions has unacknowledged implications for the conditions under which the outputs of AI can serve as reasons for belief. Following Robert Brandom, we argue that, because the assertion of a claim is an action, moral agency is a necessary condition for the giving and evaluating of reasons in discourse. Thus, the same considerations that suggest that machines cannot be held responsible for their actions suggest that they cannot be held to account for the epistemic value — or lack of value — of their outputs. If there is a responsibility gap, there is also a “testimony gap.” An under-recognised problem with the use of AI, then, is to ensure that it does not block the attribution of testimony to human beings or lead to individuals being held responsible for claims that they have not asserted. More generally, the “assertions” of machines are only capable of serving as justifications for belief or action where one or more people accept responsibility for them. (shrink)