The EU Artificial Intelligence Act (AIA) defines four risk categories for AI systems: unacceptable, high, limited, and minimal. However, it lacks a clear methodology for the assessment of these risks in concrete situations. Risks are broadly categorized based on the application areas of AI systems and ambiguous risk factors. This paper suggests a methodology for assessing AI risk magnitudes, focusing on the construction of real-world risk scenarios. To this scope, we propose to integrate the AIA with a framework developed by (...) the Intergovernmental Panel on Climate Change (IPCC) reports and related literature. This approach enables a nuanced analysis of AI risk by exploring the interplay between (a) risk determinants, (b) individual drivers of determinants, and (c) multiple risk types. We further refine the proposed methodology by applying a proportionality test to balance the competing values involved in AI risk assessment. Finally, we present three uses of this approach under the AIA: to implement the Regulation, to assess the significance of risks, and to develop internal risk management systems for AI deployers. (shrink)

As the artificial intelligence (AI) ethics field is currently working towards its operationalisation, ethics review as carried out by research ethics committees (RECs) constitutes a powerful, but so far underdeveloped, framework to make AI ethics effective in practice at the research level. This article contributes to the elaboration of research ethics frameworks for research projects developing and/or using AI. It highlights that these frameworks are still in their infancy and in need of a structure and criteria to ensure AI research (...) projects advance in a way that respects norms and principles. This article proposes to draw from the European Union’s AI Act currently in development to shape these frameworks. Although, in the current form of the draft (as of August 2023), the obligations of the AI Act do not apply to scientific research, it is most likely that they will still have a strong impact on AI research considering the need to anticipate market placement or to test new tools in real world conditions. This article investigates what the risk-based approach in the AI Act implies for research ethics and highlights some AI Act obligations of particular value to implement for ethics review processes. (shrink)

The EU proposal for the Artificial Intelligence Act (AIA) defines four risk categories: unacceptable, high, limited, and minimal. However, as these categories statically depend on broad fields of application of AI, the risk magnitude may be wrongly estimated, and the AIA may not be enforced effectively. This problem is particularly challenging when it comes to regulating general-purpose AI (GPAI), which has versatile and often unpredictable applications. Recent amendments to the compromise text, though introducing context-specific assessments, remain insufficient. To address this, (...) we propose applying the risk categories to specific AI scenarios, rather than solely to fields of application, using a risk assessment model that integrates the AIA with the risk approach arising from the Intergovernmental Panel on Climate Change (IPCC) and related literature. This model enables the estimation of the magnitude of AI risk by considering the interaction between (a) risk determinants, (b) individual drivers of determinants, and (c) multiple risk types. We use large language models (LLMs) as an example. (shrink)

Regulation is nothing without enforcement. This particularly holds for the dynamic field of emerging technologies. Hence, this article has two ambitions. First, it explains how the EU´s new Artificial Intelligence Act (AIA) will be implemented and enforced by various institutional bodies, thus clarifying the governance framework of the AIA. Second, it proposes a normative model of governance, providing recommendations to ensure uniform and coordinated execution of the AIA and the fulfilment of the legislation. Taken together, the article explores how the (...) AIA may be implemented by national and EU institutional bodies, encompassing longstanding bodies, such as the European Commission, and those newly established under the AIA, such as the AI Office. It investigates their roles across supranational and national levels, emphasizing how EU regulations influence institutional structures and operations. These regulations may not only directly dictate the structural design of institutions but also indirectly request administrative capacities needed to enforce the AIA. (shrink)

Despite recent efforts by the Artificial Intelligence (AI) community to move towards standardised procedures for documenting models, methods, systems or datasets, there is currently no methodology focused on use cases aligned with the risk-based approach of the European AI Act (AI Act). In this paper, we propose a new framework for the documentation of use cases that we call use case cards, based on the use case modelling included in the Unified Markup Language (UML) standard. Unlike other documentation methodologies, we (...) focus on the intended purpose and operational use of an AI system. It consists of two main parts: firstly, a UML-based template, tailored to allow implicitly assessing the risk level of the AI system and defining relevant requirements, and secondly, a supporting UML diagram designed to provide information about the system-user interactions and relationships. The proposed framework is the result of a co-design process involving a relevant team of EU policy experts and scientists. We have validated our proposal with 11 experts with different backgrounds and a reasonable knowledge of the AI Act as a prerequisite. We provide the 5 use case cards used in the co-design and validation process. Use case cards allows framing and contextualising use cases in an effective way, and we hope this methodology can be a useful tool for policy makers and providers for documenting use cases, assessing the risk level, adapting the different requirements and building a catalogue of existing usages of AI. (shrink)

Human oversight has become a key mechanism for the governance of artificial intelligence (“AI”). Human overseers are supposed to increase the accuracy and safety of AI systems, uphold human values, and build trust in the technology. Empirical research suggests, however, that humans are not reliable in fulfilling their oversight tasks. They may be lacking in competence or be harmfully incentivised. This creates a challenge for human oversight to be effective. In addressing this challenge, this article aims to make three contributions. (...) First, it surveys the emerging laws of oversight, most importantly the European Union’s Artificial Intelligence Act (“AIA”). It will be shown that while the AIA is concerned with the competence of human overseers, it does not provide much guidance on how to achieve effective oversight and leaves oversight obligations for AI developers underdefined. Second, this article presents a novel taxonomy of human oversight roles, differentiated along whether human intervention is constitutive to, or corrective of a decision made or supported by an AI. The taxonomy allows to propose suggestions for improving effectiveness tailored to the type of oversight in question. Third, drawing on scholarship within democratic theory, this article formulates six normative principles which institutionalise distrust in human oversight of AI. The institutionalisation of distrust has historically been practised in democratic governance. Applied for the first time to AI governance, the principles anticipate the fallibility of human overseers and seek to mitigate them at the level of institutional design. They aim to directly increase the trustworthiness of human oversight and to indirectly inspire well-placed trust in AI governance. (shrink)

In this paper, I will investigate the possible impact of weak artificial intelligence (more specifically, I will concentrate on deep learning) on human capability of action. For this goal, I will first address Arendt’s philosophy of action, which seeks to emphasize the distinguishing elements of action that set it apart from other forms of human activity. According to Arendt, action should be conceived as _praxis_, an activity that has its goal in its own very performance. The authentic meaning of action (...) includes the “passion” for articulation of one’s own individuality; I can only manifest myself as a distinct personality insofar as I introduce myself as a novel beginning to the web of human interactions demonstrating both my relevance and distinction from others. From this Arendtian standpoint, I will analyse the impact of deep learning in modern AI from two possible angles. First, I will argue that _the direct interaction_ between AI and action is impossible. Since AI operates on the principle of efficiency, it can neither suggest certain goals for action for us nor overtake their implementation because action is not guided by the instrumental need to be efficient but by the existential desire to be someone. Second, I will also analyse the possibility of _the indirect impact_ of AI on action. More specifically, I analyse neural network’s ability to circulate actions among individuals based on mathematical calculation. As I will argue, the efficiency of this circulation that surpasses human cognitive capacities can potentially organize a broader network of interaction among individuals and serve as a catalyst for the ability to act. (shrink)

Modern generative AI systems have shown the capacity to produce remarkably fluent language, prompting debates both about their semantic understanding and, less prominently, about whether they can perform speech acts. This paper addresses the latter question, focusing on assertion. We argue that to be capable of assertion, an entity must meet two requirements: it must produce outputs with descriptive functions, and it must be capable of being sanctioned by agents with which it interacts. The second requirement arises from the nature (...) of assertion as a norm-governed social practice. Pre-trained large language models that have not been subject to fine-tuning fail to meet the first requirement. Language models that have been fine-tuned for ‘groundedness’ or ‘correctness’ may meet the first requirement, but fail the second. We also consider the significance of the point that AI systems can be used to generate proxy assertions on behalf of human agents. (shrink)

With the rapid advances in science and technology, Artificial Intelligence (AI) has been developing exponentially and transforming the world in ways we could never have envisioned. Its applications...

Many artificial intelligence (AI) systems currently used for decision-making are opaque, i.e., the internal factors that determine their decisions are not fully known to people due to the systems’ computational complexity. In response to this problem, several researchers have argued that human decision-making is equally opaque and since simplifying, reason-giving explanations (rather than exhaustive causal accounts) of a decision are typically viewed as sufficient in the human case, the same should hold for algorithmic decision-making. Here, I contend that this argument (...) overlooks that human decision-making is sometimes significantly more transparent and trustworthy than algorithmic decision-making. This is because when people explain their decisions by giving reasons for them, this frequently prompts those giving the reasons to govern or regulate themselves so as to think and act in ways that confirm their reason reports. AI explanation systems lack this self-regulative feature. Overlooking it when comparing algorithmic and human decision-making can result in underestimations of the transparency of human decision-making and in the development of explainable AI that may mislead people by activating generally warranted beliefs about the regulative dimension of reason-giving. (shrink)

The increasing demand for transparency in AI has recently come under scrutiny. The question is often posted in terms of “epistemic double standards”, and whether the standards for transparency in AI ought to be higher than, or equivalent to, our standards for ordinary human reasoners. I agree that the push for increased transparency in AI deserves closer examination, and that comparing these standards to our standards of transparency for other opaque systems is an appropriate starting point. I suggest that a (...) more fruitful exploration of this question will involve a different comparison class. We routinely treat judgments made by highly trained experts in specialized fields as fair or well-grounded even though—by the nature of expert/layperson division of epistemic labor—an expert will not be able to provide an explanation of the reasoning behind these judgments that makes sense to most other people. Regardless, laypeople are thought to be acting reasonably—and ethically—in deferring to the judgments of experts that concern their areas of specialization. I suggest that we reframe our question regarding the appropriate standards of transparency in AI as one that asks when, why, and to what degree it would be ethical to accept opacity in AI. I argue that our epistemic relation to certain opaque AI technology may be relevantly similar to the layperson’s epistemic relation to the expert in certain respects, such that the successful expert/layperson division of epistemic labor can serve as a blueprint for the ethical use of opaque AI. (shrink)

The advent of Generative AI, particularly through Large Language Models (LLMs) like ChatGPT and its successors, marks a paradigm shift in the AI landscape. Advanced LLMs exhibit multimodality, handling diverse data formats, thereby broadening their application scope. However, the complexity and emergent autonomy of these models introduce challenges in predictability and legal compliance. This paper analyses the legal and regulatory implications of Generative AI and LLMs in the European Union context, focusing on liability, privacy, intellectual property, and cybersecurity. It examines (...) the adequacy of the existing and proposed EU legislation, including the Artificial Intelligence Act (AIA), in addressing the challenges posed by Generative AI in general and LLMs in particular. The paper identifies potential gaps and shortcomings in the EU legislative framework and proposes recommendations to ensure the safe and compliant deployment of generative models. (shrink)

Artificial intelligence (AI) systems are increasingly being used in healthcare, thanks to the high level of performance that these systems have proven to deliver. So far, clinical applications have focused on diagnosis and on prediction of outcomes. It is less clear in what way AI can or should support complex clinical decisions that crucially depend on patient preferences. In this paper, we focus on the ethical questions arising from the design, development and deployment of AI systems to support decision-making around (...) cardiopulmonary resuscitation and the determination of a patient’s Do Not Attempt to Resuscitate status (also known as code status). The COVID-19 pandemic has made us keenly aware of the difficulties physicians encounter when they have to act quickly in stressful situations without knowing what their patient would have wanted. We discuss the results of an interview study conducted with healthcare professionals in a university hospital aimed at understanding the status quo of resuscitation decision processes while exploring a potential role for AI systems in decision-making around code status. Our data suggest that (1) current practices are fraught with challenges such as insufficient knowledge regarding patient preferences, time pressure and personal bias guiding care considerations and (2) there is considerable openness among clinicians to consider the use of AI-based decision support. We suggest a model for how AI can contribute to improve decision-making around resuscitation and propose a set of ethically relevant preconditions—conceptual, methodological and procedural—that need to be considered in further development and implementation efforts. (shrink)

Concerns about threats to human autonomy feature prominently in the field of AI ethics. One aspect of this concern relates to the use of AI systems for problematically manipulative influence. In response to this, the European Union’s draft AI Act (AIA) includes a prohibition on AI systems deploying subliminal techniques that alter people’s behavior in ways that are reasonably likely to cause harm (Article 5(1)(a)). Critics have argued that the term ‘subliminal techniques’ is too narrow to capture the target cases (...) of AI-based manipulation. We propose a definition of ‘subliminal techniques’ that (a) is grounded on a plausible interpretation of the legal text; (b) addresses all or most of the underlying ethical concerns motivating the prohibition; (c) is defensible from a scientific and philosophical perspective; and (d) does not over-reach in ways that impose excessive administrative and regulatory burdens. The definition is meant to provide guidance for design teams seeking to pursue responsible and ethically aligned AI innovation. (shrink)

Risks connected with AI systems have become a recurrent topic in public and academic debates, and the European proposal for the AI Act explicitly adopts a risk-based tiered approach that associates different levels of regulation with different levels of risk. However, a comprehensive and general framework to think about AI-related risk is still lacking. In this work, we aim to provide an epistemological analysis of such risk building upon the existing literature on disaster risk analysis and reduction. We show how (...) a multi-component analysis of risk, that distinguishes between the dimensions of hazard, exposure, and vulnerability, allows us to better understand the sources of AI-related risks and effectively intervene to mitigate them. This multi-component analysis also turns out to be particularly useful in the case of general-purpose and experimental AI systems, for which it is often hard to perform both ex-ante and ex-post risk analyses. (shrink)

Criminal liability for acts committed by AI systems has recently become a hot legal topic. This paper includes three different contributions. The first contribution is an analysis of the extent to which an AI system can satisfy the requirements for criminal liability: accomplishing an actus reus, having the corresponding mens rea, possessing the cognitive capacities needed for responsibility. The second contribution is a discussion of criminal activity accomplished by an AI entity, with reference to a recent case involving an online (...) bot, the Random Darknet Shopper. This discussion will provide the context for the analysis of commonalities and differences between criminal activities by humans and by artificial systems. The third contribution concerns the evaluation of different ways of addressing criminal activities by AI systems in a regulatory perspective. (shrink)

Could the concept of 'evil' apply to AI? Drawing on PF Strawson's framework of reactive attitudes, this paper argues that we can understand evil as involving agents who are neither fully inside nor fully outside our moral practices. It involves agents whose abilities and capacities are enough to make them morally responsible for their actions, but whose behaviour is far enough outside of the norms of our moral practices to be labelled 'evil'. Understood as such, the paper argues that, when (...) it comes to AI and evil, we have more to fear from semi-autonomous, hybrid agents: agents who have enough of a capacity to operate within a minimal form of human practices of participation and responsibility, yet remain partially outside of it, capable of acting from outside the parameters of ordinary interactions in certain circumstances, perhaps sometimes for reasons unknown. (shrink)

There is a non-trivial chance that sometime in the future, someone will build an artificial general intelligence that will surpass human-level cognitive proficiency and go on to become “superintelligent”, vastly outperforming humans. The advent of superintelligent AI has great potential, for good or ill. It is therefore imperative that we find a way to ensure—long before one arrives—that any superintelligence we build will consistently act in ways congenial to our interests. This is a very difficult challenge in part because most (...) of the final goals we could give an AI admit of so-called “perverse instantiations”. I propose a novel solution to this puzzle: instruct the AI to love humanity. The proposal is compared with Yudkowsky’s Coherent Extrapolated Volition, and Bostrom’s Moral Modeling proposals. (shrink)

The growing number of ‘smart’ instruments, those equipped with AI, has raised concerns because these instruments make autonomous decisions; that is, they act beyond the guidelines provided them by programmers. Hence, the question the makers and users of smart instrument face is how to ensure that these instruments will not engage in unethical conduct. The article suggests that to proceed we need a new kind of AI program—oversight programs—that will monitor, audit, and hold operational AI programs accountable.

Artificial intelligence and automation have become pervasive in news media, influencing journalism from news gathering to news distribution. As algorithms are increasingly determining editorial decisions, specific concerns have been raised with regard to the responsible and accountable use of AI-driven tools by news media, encompassing new regulatory and ethical questions. This contribution aims to analyze whether and to what extent the use of AI technology in news media and journalism is currently regulated and debated within the European Union and the (...) Council of Europe. Through a document analysis of official policy documents, combined with a data mining approach and an inductive thematic analysis, the study looks at how news media are dealt with, in particular regarding their responsibilities towards their users and society. The findings show that regulatory frameworks about AI rarely include media, but if they do, they associate them with issues such as disinformation, data, and AI literacy, as well as diversity, plurality, and social responsibility. (shrink)

A simple argument suggests that we can fruitfully model advanced AI systems using expected utility theory. According to this argument, an agent will need to act as if maximising expected utility if they’re to avoid exploitation. Insofar as we should expect advanced AI to avoid exploitation, it follows that we should expected advanced AI to act as if maximising expected utility. I spell out this argument more carefully and demonstrate that it fails, but show that the manner of its failure (...) is instructive: in exploring the argument, we gain insight into how to model advanced AI systems. (shrink)

Despite pronouncements about the inevitable diffusion of artificial intelligence and autonomous technologies, in practice, it is human behavior, not technology in a vacuum, that dictates how technology seeps into—and changes—societies. To better understand how human preferences shape technological adoption and the spread of AI-enabled autonomous technologies, we look at representative adult samples of US public opinion in 2018 and 2020 on the use of four types of autonomous technologies: vehicles, surgery, weapons, and cyber defense. By focusing on these four diverse (...) uses of AI-enabled autonomy that span transportation, medicine, and national security, we exploit the inherent variation between these AI-enabled autonomous use cases. We find that those with familiarity and expertise with AI and similar technologies were more likely to support all of the autonomous applications we tested (except weapons) than those with a limited understanding of the technology. Individuals that had already delegated the act of driving using ride-share apps were also more positive about autonomous vehicles. However, familiarity cut both ways; individuals are also less likely to support AI-enabled technologies when applied directly to their life, especially if technology automates tasks they are already familiar with operating. Finally, we find that familiarity plays little role in support for AI-enabled military applications, for which opposition has slightly increased over time. (shrink)

In this paper, we address the question of whether AI should be used for suicide prevention on social media data. We focus on algorithms that can identify persons with suicidal ideation based on their postings on social media platforms and investigate whether private companies like Facebook are justified in using these. To find out if that is the case, we start with providing two examples for AI-based means of suicide prevention in social media. Subsequently, we frame suicide prevention as an (...) issue of beneficence, develop two fictional cases to explore the scope of the principle of beneficence and apply the lessons learned to Facebook’s employment of AI for suicide prevention. We show that Facebook is neither acting under an obligation of beneficence nor acting meritoriously. This insight leads us to the general question of who is entitled to help. We conclude that private companies like Facebook can play an important role in suicide prevention, if they comply with specific rules which we derive from beneficence and autonomy as core principles of biomedical ethics. At the same time, public bodies have an obligation to create appropriate framework conditions for AI-based tools of suicide prevention. As an outlook we depict how cooperation between public and private institutions can make an important contribution to combating suicide and, in this way, put the principle of beneficence into practice. (shrink)

Developers and vendors of large language models (“LLMs”) — such as ChatGPT, Google Bard, and Microsoft’s Bing at the forefront—can be subject to Health Insurance Portability and Accountability Act of 1996 (“HIPAA”) when they process protected health information (“PHI”) on behalf of the HIPAA covered entities. In doing so, they become business associates or subcontractors of a business associate under HIPAA.

Using speech act theory from the Philosophy of Language, this paper attempts to develop an ethical framework for the phenomenon of speech processing. We use the concepts of the illocutionary force and the illocutionary content of a speech act to explain the ethics of speech processing. By emphasizing the different stages involved in speech processing, we explore the distinct ethical issues that arise in relation to each stage. Input, processing, and output are the different ethically relevant stages under which a (...) spoken item or a speech navigates within the range of speech-processing modules. Employing the illocutionary force-content distinction, we specify and characterize the input-related ethical issues, the output-related ethical issues, and the processing-related ethical issues involved in speech processing. Together with illocutionary force-content distinction, we employ the data-information distinction to characterize the stage-wise ethical issues in the phenomenon of speech processing as the ethics of collecting (speech) data, the ethics of contextualizing (speech) data/information, and the ethics of releasing the con-textualized information (processed speech). Immediate ethical issues that arise from the range of speech processing modules are distinguished from distant ethical issues. We also indicate the nature of ethical issues that arise from Speaker Independent speech technologies. (shrink)

Adolf Reinach’s theory of social acts and Czesław Znamierowski theory of the environment can show a new perspective of analysis in the fields of computer science and digital communication. This paper will begin analysing the performance of social acts in two categories of digital environments: (i) fictional digital environment and (ii) real digital environment. The analysis will be supported by examples from the history of computer science. In both kinds of digital environments, organigrams play a significant role and depend on (...) the users’ digital power to perform a real or fictional social act. Finally, the paper will analyse one of the possible roles that AI plays in performing social acts in digital environments. It will show how AI could affect the perception of social acts. (shrink)

We paraphrase Descartes’ famous dictum in the area of AI ethics where the “I doubt and therefore I am” is suggested as a necessary aspect of morality. Therefore AI, which cannot doubt itself, cannot possess moral agency. Of course, this is not the end of the story. We explore various aspects of the human mind that substantially differ from AI, which includes the sensory grounding of our knowing, the act of understanding, and the significance of being able to doubt ourselves. (...) The foundation of our argument is the discipline of ethics, one of the oldest and largest knowledge projects of human history, yet, we seem only to be beginning to get a grasp of it. After a couple of thousand years of studying the ethics of humans, we (humans) arrived at a point where moral psychology suggests that our moral decisions are intuitive, and all the models from ethics become relevant only when we explain ourselves. This recognition has a major impact on what and how we can do regarding AI ethics. We do not offer a solution, we explore some ideas and leave the problem open, but we hope somewhat better understood than before our study. (shrink)

It has become fashionable to worry about the development of superintelligent AI that results in the destruction of humanity. This worry is not without merit; but it may be overstated. This paper explores some previously undiscussed reasons to be optimistic that; even if superintelligent AI does arise; it will not destroy us. These have to do with the possibility that a superintelligent AI will become mired in skeptical worries that its superintelligence cannot help it to solve. I argue that superintelligent (...) AIs may lack the psychological idiosyncracies that allow humans to act in the face of skeptical problems; and so as a result they may become paralyzed in the face of these problems in a way that humans are not. (shrink)

Artificial intelligence increasingly executes tasks that previously only humans could do, such as drive a car, fight in war, or perform a medical operation. However, as the very best AI systems tend to be the least controllable and the least transparent, some scholars argued that humans can no longer be morally responsible for some of the AI-caused outcomes, which would then result in a responsibility gap. In this paper, I assume, for the sake of argument, that at least some of (...) the most sophisticated AI systems do indeed create responsibility gaps, and I ask whether we can bridge these gaps at will, viz. whether certain people could take responsibility for AI-caused harm simply by performing a certain speech act, just as people can give permission for something simply by performing the act of consent. So understood, taking responsibility would be a genuine normative power. I first discuss and reject the view of Champagne and Tonkens, who advocate a view of taking liability. According to this view, a military commander can and must, ahead of time, accept liability to blame and punishment for any harm caused by autonomous weapon systems under her command. I then defend my own proposal of taking answerability, viz. the view that people can makes themselves morally answerable for the harm caused by AI systems, not only ahead of time but also when harm has already been caused. (shrink)

The book examines the extent to which Chinese cyber and network security laws and policies act as a constraint on the emergence of Chinese entrepreneurialism and innovation. Specifically, how the contradictions and tensions between data localisation laws affect innovation in artificial intelligence. The book surveys the globalised R&D networks, and how the increasing use of open-source platforms by leading Chinese AI firms during 2017–2020, exacerbated the apparent contradiction between Network Sovereignty and Chinese innovation. The drafting of the Cyber Security Law (...) did not anticipate the changing nature of globalised AI innovation. It is argued that the deliberate deployment of what the book refers to as 'fuzzy logic' in drafting the Cyber Security Law allowed regulators to subsequently interpret key terms regarding data in that Law in a fluid and flexible fashion to benefit Chinese innovation. (shrink)

The emergence of AI is posing serious challenges to standard conceptions of moral status. New non-biological entities are able to act and make decisions rationally. The question arises, in this regard, as to whether AI systems possess or can possess the necessary properties to be morally considerable. In this chapter, we have undertaken a systematic analysis of the various debates that are taking place about the moral status of AI. First, we have discussed the possibility that AI systems, by virtue (...) of its new agential capabilities, can be understood as a moral agent. Discussions between those defending mentalist and anti-mentalist positions have revealed many nuances and particularly relevant theoretical aspects. Second, given that an AI system can hardly be an entity qualified to be responsible, we have delved into the responsibility gap and the different ways of understanding and addressing it. Third, we have provided an overview of the current and potential patientist capabilities that AI systems possess. This has led us to analyze the possibilities of AI possessing moral patiency. In addition, we have addressed the question of the moral and legal rights of AI. Finally, we have introduced the two most relevant authors of the relational turn on the moral status of AI, Mark Coeckelbergh and David Gunkel, who have been led to defend a relational approach to moral life as a result of the problems associated with the ontological understanding of moral status. (shrink)

In this paper, we propose a methodology to maximize the benefits of interdisciplinary cooperation in AI research groups. Firstly, we build the case for the importance of interdisciplinarity in research groups as the best means to tackle the social implications brought about by AI systems, against the backdrop of the EU Commission proposal for an Artificial Intelligence Act. As we are an interdisciplinary group, we address the multi-faceted implications of the mass-scale diffusion of AI-driven technologies. The result of our exercise (...) lead us to postulate the necessity of a behavioural theory that standardizes the interaction process of interdisciplinary groups. In light of this, we conduct a review of the existing approaches to interdisciplinary research on AI appliances, leading to the development of methodologies like ethics-by-design and value-sensitive design, evaluating their strengths and weaknesses. We then put forth an iterative process theory hinging on a narrative approach consisting of four phases: definition of the hypothesis space, building-up of a common lexicon, scenario-building, interdisciplinary self-assessment. Finally, we identify the most relevant fields of application for such a methodology and discuss possible case studies. (shrink)

There is a non-trivial chance that sometime in the (perhaps somewhat distant) future, someone will build an artificial general intelligence that will surpass human-level cognitive proficiency and go on to become "superintelligent", vastly outperforming humans. The advent of superintelligent AI has great potential, for good or ill. It is therefore imperative that we find a way to ensure-long before one arrives-that any superintelligence we build will consistently act in ways congenial to our interests. This is a very difficult challenge in (...) part because most of the final goals we could give an AI admit of so-called "perverse instantiations". I propose a novel solution to this puzzle: instruct the AI to love humanity. The proposal is compared with Yudkowsky's Coherent Extrapolated Volition, and Bostrom's Moral Modeling proposals. (shrink)

In recent years, the European Union has advanced towards responsible and sustainable Artificial Intelligence (AI) research, development and innovation. While the Ethics Guidelines for Trustworthy AI released in 2019 and the AI Act in 2021 set the starting point for a European Ethical AI, there are still several challenges to translate such advances into the public debate, education and practical learning. This paper contributes towards closing this gap by reviewing the approaches that can be found in the existing literature and (...) by interviewing 11 experts across five countries to help define educational strategies, competencies and resources needed for the successful implementation of Trustworthy AI in Higher Education (HE) and to reach students from all disciplines. The findings are presented in the form of recommendations both for educators and policy incentives, translating the guidelines into HE teaching and practice, so that the next generation of young people can contribute to an ethical, safe and cutting-edge AI made in Europe. (shrink)

On 21 April 2021, the European Commission published the proposal of the new EU Artificial Intelligence Act (AIA) — one of the most influential steps taken so far to regulate AI internationally. This article highlights some foundational aspects of the Act and analyses the philosophy behind its proposal.

With the widespread use of artificial intelligence, it becomes crucial to provide information about these systems and how they are used. Governments aim to disclose their use of algorithms to establish legitimacy and the EU AI Act mandates forms of transparency for all high-risk and limited-risk systems. Yet, what should the standards for transparency be? What information is needed to show to a wide public that a certain system can be used legitimately and responsibly? I argue that process-based approaches fail (...) to satisfy, as knowledge about the development process is insufficient to predict the properties of the resulting system. Current outcome-based approaches [Mitchell et al., 2019; Loi et al., 2021] are also criticized for a lack of attention to the broader socio-technical system and failure to account for empirical results that show that people care about more than just the outcomes of a process [as reported by Meyerson et al. (Procedural justice and relational theory: Empirical, philosophical, and legal perspectives, Taylor & Francis, 2021)]. Instead, I propose value-based transparency, on which the information we need to provide is what values have been considered in the design and how successful these have been realized in the final system. This can handle the objections to other frameworks, matches with current best practices on the design of responsible AI and provides the public with information on the crucial aspects of a system’s design. (shrink)

The use of black box algorithms in medicine has raised scholarly concerns due to their opaqueness and lack of trustworthiness. Concerns about potential bias, accountability and responsibility, patient autonomy and compromised trust transpire with black box algorithms. These worries connect epistemic concerns with normative issues. In this paper, we outline that black box algorithms are less problematic for epistemic reasons than many scholars seem to believe. By outlining that more transparency in algorithms is not always necessary, and by explaining that (...) computational processes are indeed methodologically opaque to humans, we argue that the reliability of algorithms provides reasons for trusting the outcomes of medical artificial intelligence. To this end, we explain how computational reliabilism, which does not require transparency and supports the reliability of algorithms, justifies the belief that results of medical AI are to be trusted. We also argue that several ethical concerns remain with black box algorithms, even when the results are trustworthy. Having justified knowledge from reliable indicators is, therefore, necessary but not sufficient for normatively justifying physicians to act. This means that deliberation about the results of reliable algorithms is required to find out what is a desirable action. Thus understood, we argue that such challenges should not dismiss the use of black box algorithms altogether but should inform the way in which these algorithms are designed and implemented. When physicians are trained to acquire the necessary skills and expertise, and collaborate with medical informatics and data scientists, black box algorithms can contribute to improving medical care. (shrink)

The use of artificial intelligence (AI) in hiring entails vast ethical challenges. As such, using an ethical lens to study this phenomenon is to better understand whether and how AI matters in hiring. In this paper, we examine whether ethical perceptions of using AI in the hiring process influence individuals’ trust in the organizations that use it. Building on the organizational trust model and the unified theory of acceptance and use of technology, we explore whether ethical perceptions are shaped by (...) individual differences in performance expectancy and social influence and how they, in turn, impact organizational trust. We collected primary data from over 300 individuals who were either active job seekers or who had recent hiring experience to capture perceptions across the full range of hiring methods. Our findings indicate that performance expectancy, but not social influence, impacts the ethical perceptions of AI in hiring, which in turn influence organizational trust. Additional analyses indicate that these findings vary depending on the type of hiring methods AI is used for, as well as on whether participants are job seekers or individuals with hiring experience. Our study offers theoretical and practical implications for ethics in HRM and informs policy implementation about when and how to use AI in hiring methods, especially as it pertains to acting ethically and trustworthily. (shrink)

Reasoning about causation in fact is an essential element of attributing legal responsibility. Therefore, the automation of the attribution of legal responsibility requires a modelling effort aimed at the following: a thorough understanding of the relation between the legal concepts of responsibility and of causation in fact; a thorough understanding of the relation between causation in fact and the common sense concept of causation; and, finally, the specification of an ontology of the concepts that are minimally required for (automatic) common (...) sense reasoning about causation. This article offers a worked-out example of the indicated analysis. Such example consists of: a definition of the legal concept of responsibility (in terms of liability and accountability); a definition of the legal concept of causation in fact (in terms of the initiation of physical processes by an agent and of the provision of reasons and/or opportunities to other agents); CausatiOnt, an AI-like ontology of the common sense (causal) concepts that are minimally needed for reasoning about the legal concept of causation in fact (in particular, the concepts of category, dimension, object, agent, process, event and act). (shrink)

With the rapid advancement of AI, there exists a possibility of rogue human actor(s) taking control of a potent AI system or an AI system redefining its objective function such that it presents an existential threat to mankind or severely curtails its freedom. Therefore, some suggest an outright ban on AI development while others profess international agreement on constraining specific types of AI. These approaches are untenable because countries will continue developing AI for national defense, regardless. Some suggest having an (...) all-powerful benevolent one-AI that will act as an AI nanny. However, such an approach relies on the everlasting benevolence of one-AI, an untenable proposition. Furthermore, such an AI is itself subject to capture by a rogue actor. We present an alternative approach that uses existing mechanisms and time-tested economic concepts of competition and marginal analysis to limit centralization and integration of AI, rather than AI itself. Instead of depending on international consensus it relies on countries working in their best interests. We recommend that through regulation and subsidies countries promote independent development of competing AI technologies, especially those with decentralized architecture. The Sherman Antitrust Act can be used to limit the domain of an AI system, training module, or any of its components. This will increase the segmentation of potent AI systems and force technological incompatibility across systems. Finally, cross-border communication between AI-enabled systems should be restricted, something countries like China and the US are already inclined to do to serve their national interests. Our approach can ensure the availability of numerous sufficiently powerful AI systems largely disconnected from each other that can be called upon to identify and neutralize rogue systems when needed. This setup can provide sufficient deterrence to any rational human or AI system from attempting to exert undue control. (shrink)

Modelling that exploits visual elements and information visualisation are important areas that have contributed immensely to understanding and the computerisation advancements in many domains and yet remain unexplored for the benefit of the law and legal practice. This paper investigates the challenge of modelling and expressing structures and processes in legislation and the law by using visual modelling and information visualisation (InfoVis) to assist accessibility of legal knowledge, practice and knowledge formalisation as a basis for legal AI. The paper uses (...) a subset of the well-defined Unified Modelling Language (UML) to visually express the structure and process of the legislation and the law to create visual flow diagrams called lawmaps, which form the basis of further formalisation. A lawmap development methodology is presented and evaluated by creating a set of lawmaps for the practice of conveyancing and the Landlords and Tenants Act 1954 of the United Kingdom. This paper is the first of a new breed of preliminary solutions capable of application across all aspects, from legislation to practice; and capable of accelerating development of legal AI. (shrink)

This article argues that large language models (LLMs) should be interpreted as a form of gods. In a theological sense, a god is an immortal being that exists beyond time and space. This is clearly nothing like LLMs. In an anthropological sense, however, a god is rather defined as the personified authority of a group through time—a conceptual tool that molds a collective of ancestors into a unified agent or voice. This is exactly what LLMs are. They are products of (...) vast volumes of data, literally traces of past human (speech) acts, synthesized into a single agency that is (falsely) experienced by users as extra-human. This reconceptualization, I argue, opens up new avenues of critique of LLMs by allowing the mobilization of theoretical resources from centuries of religious critique. For illustration, I draw on the Marxian religious philosophy of Martin Hägglund. From this perspective, the danger of LLMs emerge not only as bias or unpredictability, but as a temptation to abdicate our spiritual and ultimately democratic freedom in favor of what I call a tyranny of the past. (shrink)

In lieu of an abstract, here is a brief excerpt of the content:The Touching TestAI and the Future of Human IntimacyMartha J. Reineke (bio)Each Friday, the New York Times publishes Love Letters, a compendium of articles on courtship. A recent story featured Melinda, a real estate agent, and Calvin, a human resources director.1 They had met at a market deli counter. On their first date, a lasagna dinner at Melinda's home, Calvin posed the question, "What are you looking for in (...) a man?" To his surprise, Melinda reached into her pocket and pulled out a two-page list, "Melinda's Husband Traits." Her 46-item list begins with "Christian, honor God, put Him first" and includes as entry number 37, "He likes to be touched and likes to touch me." Notwithstanding laudable features of Melinda's inventory, if David Levy, President of the International Computer Games Association and an expert on artificial intelligence (AI), is correct in his predictions, in the years ahead, women like Melinda who know exactly what they are looking for in a husband will desire something radically new: a robot. After all, artificial agents will soon possess or improve on every quality the romantically inclined previously have sought in a human partner. [End Page 123]Levi predicts robots will be preferred romantic partners by the middle of this century.2 Capable of providing erotic satisfaction and creating deep emotional bonds with human partners, these robots will be an achievable outcome of AI engineering that already has brought to market industrial and service robots, virtual pets, and caregiver robots for the elderly. Levi offers a research-based inventory of traits a robot will offer its human partner: kindness and understanding, a sense of humor, expressiveness, intelligence, and sociability.3 He pinpoints additional features that will facilitate romance: The ideal robot partner will elicit erotic attraction, give evidence of readiness for a serious relationship with its verbal assertions, and possess a capacity for mutual intimacy.4 Levi's argument is intentionally provocative, for he emphasizes the centrality of sex to future robot/human pairings, waxing rhapsodic about the performative superiority of robots to humans in the sexual act.5Elaborating on potential human love relationships with robots, Levi asserts that a robot will need to master imitation if a human is to want it as a romantic partner. Not only is imitation a central aspect of every human-to-human relationship, but also imitation has been identified by researchers as a crucial element in romantic desire.6 Levi states that human-to-human attraction is founded on similarities between prospective partners in appearance, personality, and social context. It is sustained only through what Levi calls "reciprocal liking."7 For romantic relationships between humans and robots to flourish, they too must engage in acts of reciprocal imitation.In highlighting the necessity of imitation in effective robot and human interactions, Levi draws close to René Girard's theory of "mimetic desire." According to Girard, imitation structures all "wanting,"8 delineating what we aspire to look like or hope to possess and determining what we value, take comfort in, or experience as a threat. Indeed, the individual is never the origin point of its desires. Each of us always desires mimetically—according to the desire of the other.9 As Scott Garrels confirms in Mimesis and Science, researchers studying imitation in developmental, comparative, and social psychology, as well as in neurophysiology and cognitive neuroscience, support Girard's claims: Imitation is fundamental in human life.10 Likewise, Girard concurs with Levi that romantic relationships are primary examples of imitative desire. As described by Girard in Deceit, Desire, and the Novel, imitative behaviors form the romantic triangles that feature prominently in fiction: for example, Odette and the many men enchanted by her in Proust's In Search of Lost Time; Anselmo, Lotario, and Camilla in Don Quixote; and Velchaninov and Troussotzki, along with Troussotzki's deceased wife and prospective fiancée, in Dostoyevsky's The Eternal Husband.11 [End Page 124]Notwithstanding similarities in Levi and Girard's views of imitation, they differ in a crucial respect. Only Girard acknowledges just how extraordinarily fraught mimesis can become. When imitators compete (e.g., for things, status, and experiences), they will contest... (shrink)

There is widespread agreement that there should be a principle requiring that artificial intelligence be ‘explicable’. Microsoft, Google, the World Economic Forum, the draft AI ethics guidelines for the EU commission, etc. all include a principle for AI that falls under the umbrella of ‘explicability’. Roughly, the principle states that “for AI to promote and not constrain human autonomy, our ‘decision about who should decide’ must be informed by knowledge of how AI would act instead of us” :689–707, 2018). There (...) is a strong intuition that if an algorithm decides, for example, whether to give someone a loan, then that algorithm should be explicable. I argue here, however, that such a principle is misdirected. The property of requiring explicability should attach to a particular action or decision rather than the entity making that decision. It is the context and the potential harm resulting from decisions that drive the moral need for explicability—not the process by which decisions are reached. Related to this is the fact that AI is used for many low-risk purposes for which it would be unnecessary to require that it be explicable. A principle requiring explicability would prevent us from reaping the benefits of AI used in these situations. Finally, the explanations given by explicable AI are only fruitful if we already know which considerations are acceptable for the decision at hand. If we already have these considerations, then there is no need to use contemporary AI algorithms because standard automation would be available. In other words, a principle of explicability for AI makes the use of AI redundant. (shrink)

In this paper, we demonstrate how the language and reasonings that academics, developers, consumers, marketers, and journalists deploy to accept or reject AI as authentic intelligence has far-reaching bearing on how we understand our human intelligence and condition. The discourse on AI is part of what we call the “authenticity negotiation process” through which AI’s “intelligence” is given a particular meaning and value. This has implications for scientific theory, research directions, ethical guidelines, design principles, funding, media attention, and the way (...) people relate to and act upon AI. It also has great impact on humanity’s self-image and the way we negotiate what it means to be human, existentially, culturally, politically, and legally. We use a discourse analysis of academic papers, AI education programs, and online discussions to demonstrate how AI itself, as well as the products, services, and decisions delivered by AI systems are negotiated as authentic or inauthentic intelligence. In this negotiation process, AI stakeholders indirectly define and essentialize what being human(like) means. The main argument we will develop is that this process of indirectly defining and essentializing humans results in an elimination of the space for humans to be indeterminate. By eliminating this space and, hence, denying indeterminacy, the existential condition of the human being is jeopardized. Rather than re-creating humanity in AI, the AI discourse is re-defining what it means to be human and how humanity is valued and should be treated. (shrink)

The development of machine learning and deep learning (DL) in the field of AI (artificial intelligence) is the direct result of the advancement of nano-electronics. Machine learning is a function that provides the system with the capacity to learn from data without being programmed explicitly. It is basically a mathematical and probabilistic model. DL is part of machine learning methods based on artificial neural networks, simply called neural networks (NNs), as they are inspired by the biological NNs that constitute organic (...) brains. Despite its similarity to biological organs such as human brains, major problems arise in trying to attribute moral responsibility to autonomic systems based on hardware including nano-electronic devices, which are sought to replace humans (moral agents) in the context of AI. It is suggested that the required emotional environment which enables actions according to reasons in humans is not witnessed at AI devices. Though AI technology raises an enticing resemblance to human actions, this resemblance is to be considered with a skeptical eye. It is because actions are associated with reasons while causes are connected to operations. Human agents are capable of acting upon their reasons, while AI devices are limited only to operations as they are conditioned by their programming, which is considered as an embodiment of some causes. As moral responsibility goes hand in hand with the capacity to act (according to reasons), attributing moral responsibility to AI devices is revealed to be but a misleading metaphor. (shrink)

The creation of machines with intelligence comparable to human beings—so-called "human-level” and “general” intelligence—is often regarded as the Holy Grail of Artificial Intelligence (AI) research. However, many prominent discussions of AI lean heavily on the notion of human-level intelligence to frame AI research, but then rely on conceptions of human cognitive capacities, including “common sense,” that are sketchy, one-sided, philosophically loaded, and highly contestable. Our goal in this essay is to bring into view some underappreciated features of the practical intelligence (...) involved in ordinary human agency. These features of practical intelligence are implicit in the structure of our first-person experience of embodied and situated agency, deliberation, and human interaction. We argue that spelling out these features and their implications reveals a fundamental distinction between two forms of intelligence in action, or what we call “efficient task-completion” versus “intelligent engagement in activity.” This distinction helps us to see what is missing from some widely accepted ways of thinking about human-level intelligence in AI, and how human common sense is actually tied, conceptually, to the ideal of practical wisdom, or good (normative) judgment about how to act and live well. Finally, our analysis, if sound, also has implications for the important ethical question of what it means to have AI systems that are aligned with human values, or the so-called “value alignment” problem for artificial intelligence. (shrink)

The German Act on Autonomous Driving constitutes the first national framework on level four autonomous vehicles and has received attention from policy makers, AI ethics scholars and legal experts in autonomous driving. Owing to Germany’s role as a global hub for car manufacturing, the following paper sheds light on the act’s position within the ethical discourse and how it reconfigures the balance between legislation and ethical frameworks. Specifically, in this paper, we highlight areas that need to be more worked out (...) in the future either through ethical conventions, corporate measures or legal measures and examine how the law can be incorporated into the existing discourse on the regulation of technologies. Based on this examination, we derive implications for future discourse and elaborate on companies’ responsibilities in developing autonomous driving technologies in an ethical sense. (shrink)

Creativity has a special role in enabling humans to develop beyond the fulfilment of simple primary functions. This factor is significant for Artificial Intelligence (AI) developers who take replication to be the primary goal, since moves toward creating autonomous artificial-beings beg questions about their potential for creativity. Using Wittgenstein’s remarks on rule-following and language-games, I argue that although some AI programs appear creative, to call these programmed acts creative in our terms is to misunderstand the use of this word in (...) language. I conclude that replication is not the best way forward for AI development in matters of creativity. (shrink)