Andrea Aler Tubella

In this chapter, we will discuss how to incorporate values into intelligent systems by presenting literature and state-of-the-art tools around three key challenges: i) recognizing which and whose values should be considered; ii) operationalizing values and actively including them into an intelligent system’s behavior; and iii) ensuring that not only the resulting system, but the design and deployment processes, as a whole, adhere to desirable values. When it comes to “AI at work,” each of these aspects is key for its advancement. Firstly, identifying relevant stakeholders and their values is fundamental for the adoption of modern technology into any organizational process. Secondly, operationalizing values into specific constraints and behaviors of intelligent systems increases trust, rendering interactions more efficient and safer, therefore helping these tools to fulfil their function seamlessly and empowering workers with the knowledge needed to adequately operate these systems. Finally, ensuring that design and deployment processes follow values such as transparency and responsibility is of great benefit for organizations and workers, allowing them to show due diligence, compliance to governance constraints, and auditability, which clarifies the allocation of responsibility and prevents disuse.

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.

It is becoming more common to replace or augment human-based decisions with algorithmic calculations and evaluations using artificial intelligence (AI). Facial analysis systems (FA) are examples of how AI in particular is intertwined with both the most mundane and the most critical aspects of human life. Exploring images for the purposes of face detection, recognition and/or classification, FA shows an entanglement between human identity, self-presentation and computation. In this chapter, we discuss automated facial analysis technology from a queer theoretical standpoint, focusing on the concerns and risks when systems like FA are used in a binary way to categorize, measure and make decisions based on computerized assumptions about gender and sexuality. Further, we discuss issues of privacy, bias and fairness related to FA technology as well as potential improvements, for example, by using participatory design. Finally, this chapter suggests that a queer perspective on FA can create new ways to relate to technology.

Fairness is central to the ethical and responsible development and use of AI systems, with a large number of frameworks and formal notions of algorithmic fairness being available. However, many of the fairness solutions proposed revolve around technical considerations and not the needs of and consequences for the most impacted communities. We therefore want to take the focus away from definitions and allow for the inclusion of societal and relational aspects to represent how the effects of AI systems impact and are experienced by individuals and social groups. In this paper, we do this by means of proposing the ACROCPoLis framework to represent allocation processes with a modeling emphasis on fairness aspects. The framework provides a shared vocabulary in which the factors relevant to fairness assessments for different situations and procedures are made explicit, as well as their interrelationships. This enables us to compare analogous situations, to highlight the differences in dissimilar situations, and to capture differing interpretations of the same situation by different stakeholders.

Fairness is central to the ethical and responsible development and use of AI systems, with a large number of frameworks and formal notions of algorithmic fairness being available. However, many of the fairness solutions proposed revolve around technical considerations and not the needs of and consequences for the most impacted communities. We therefore want to take the focus away from definitions and allow for the inclusion of societal and relational aspects to represent how the effects of AI systems impact and are experienced by individuals and social groups. In this paper, we do this by means of proposing the ACROCPoLis framework to represent allocation processes with a modeling emphasis on fairness aspects. The framework provides a shared vocabulary in which the factors relevant to fairness assessments for different situations and procedures are made explicit, as well as their interrelationships. This enables us to compare analogous situations, to highlight the differences in dissimilar situations, and to capture differing interpretations of the same situation by different stakeholders.

The importance of fairness in machine learning models is widely acknowledged, and ongoing academic debate revolves around how to determine the appropriate fairness definition, and how to tackle the trade-off between fairness and model performance. In this paper we argue that besides these concerns, there can be ethical implications behind seemingly purely technical choices in fairness interventions in a typical model development pipeline. As an example we show that the technical choice between in-processing and post-processing is not necessarily value-free and may have serious implications in terms of who will be affected by the specific fairness intervention. The paper reveals how assessing the technical choices in terms of their ethical consequences can contribute to the design of fair models and to the related societal discussions.

Accountability is a value often mentioned in the debate on intelligent systems and their increased pervasiveness in our society. When focusing specifically on autonomous systems, a critical gap emerges: although there is much work on governance and attribution of accountability, there is a significant lack of methods for the operationalisation of accountability within the socio-technical layer of autonomous systems. In the case of autonomous unmanned aerial vehicles or drones—the critical question of how to maintain accountability as they undertake fully autonomous flights becomes increasingly important as their uses multiply in both the commercial and military fields. In this paper, we aim to fill the operationalisation gap by proposing a socio-technical framework to guarantee human oversight and accountability in drone deployments, showing its enforceability in the real case of military surveillance drones. By keeping a focus on accountability and human oversight as values, we align with the emphasis placed on human responsibility, while requiring a concretisation of what these principles mean for each specific application, connecting them with concrete socio-technical requirements. In addition, by constraining the framework to observable elements of pre- and post-deployment, we do not rely on assumptions made on the internal workings of the drone nor the technical fluency of the operator.

Artificial Intelligence (AI) applications are being used to predict and assess behaviour in multiple domains which directly affect human well-being. However, if AI is to improve people’s lives, then people must be able to trust it, by being able to understand what the system is doing and why. Although transparency is often seen as the requirement in this case, realistically it might not always be possible, whereas the need to ensure that the system operates within set moral bounds remains. In this paper, we present an approach to evaluate the moral bounds of an AI system based on the monitoring of its inputs and outputs. We place a ‘Glass-Box’ around the system by mapping moral values into explicit verifiable norms that constrain inputs and outputs, in such a way that if these remain within the box we can guarantee that the system adheres to the value. The focus on inputs and outputs allows for the verification and comparison of vastly different intelligent systems; from deep neural networks to agent-based systems. The explicit transformation of abstract moral values into concrete norms brings great benefits in terms of explainability; stakeholders know exactly how the system is interpreting and employing relevant abstract moral human values and calibrate their trust accordingly. Moreover, by operating at a higher level we can check the compliance of the system with different interpretations of the same value.

This article presents the first in a series of results that allow us to develop a theory providing finer control over the complexity of normalization, and in particular of cut elimination. By considering atoms as self-dual noncommutative connectives, we are able to classify a vast class of inference rules in a uniform and very simple way. This allows us to define simple conditions that are easily verifiable and that ensure normalization and cut elimination by way of a general theorem. In this article, we define and consider splittable systems , which essentially make up a large class of linear logics, including Multiplicative Linear Logic and BV, and we prove for them a splitting theorem , guaranteeing cut elimination and other admissibility results as corollaries. In articles to follow, we will extend this result to nonlinear logics. The final outcome will be a comprehensive theory giving a uniform treatment for most existing logics and providing a blueprint for the design of future proof systems.