Summary. We measure trust and trustworthiness in British society with a newly designed experiment using real monetary rewards and a sample of the British population. The study also asks the typical survey question that aims to measure trust, showing that it does not predict ‘trust’ as measured in the experiment. Overall, about 40% of people were willing to trust a stranger in our experiment, and their trust was rewarded half of the time. Analysis of variation in the trust behaviour in our survey suggests that trusting is more likely if people are older, their financial situation is either ‘comfortable’ or ‘difficult’ compared with ‘doing alright’ or ‘just getting by’, they are a homeowner or they are divorced, separated or never married compared with those who are married or cohabiting. Trustworthiness also is more likely among subjects who are divorced or separated relative to those who are married or cohabiting, and less likely among subjects who perceive their financial situation as ‘just getting by’ or ‘difficult’. We also analyse the effect of attitudes towards risks on trust.

Measuring people's trust

This paper shows that the law, in subtle ways, may set hitherto unrecognized incentives for the adoption of explainable machine learning applications. In doing so, we make two novel contributions. First, on the legal side, we show that to avoid liability, professional actors, such as doctors and managers, may soon be legally compelled to use explainable ML models. We argue that the importance of explainability reaches far beyond data protection law, and crucially influences questions of contractual and tort liability for the use of ML models. To this effect, we conduct two legal case studies, in medical and corporate merger applications of ML. As a second contribution, we discuss the (legally required) trade-off between accuracy and explainability and demonstrate the effect in a technical case study in the context of spam classification.

/pdf/explainable-ai-under-contract-and-tort-law-legal-incentives-34m9c3rufr.pdf

Explainable AI under contract and tort law: legal incentives and technical challenges

The alleged opacity of AI has become a major political issue over the past few years. Opening the black box, so it is argued, is indispensable to identify encroachments on user privacy, to detect biases and to prevent other potential harms. However, what is less clear is how the call for AI transparency can be translated into reasonable regulation. This Chapter argues that designing AI transparency regulation is less difficult than oftentimes assumed. Regulators profit from the fact that the legal system has already gained considerable experience with the question of how to shed light on partially opaque decision-making systems—human decisions. This experience provides lawyers with a realistic perspective of the functions of potential AI transparency legislation as well as with a set of legal instruments which can be employed to this end.

Artificial Intelligence and Transparency: Opening the Black Box

Algorithms, including artificial intelligence, are used in a variety of ways to differentiate people, services, products or positions. This study uses examples to illustrate the technical and organisational causes of discrimination risks and analyses the resulting forms of discrimination. Its particular focus is on the social risks from algorithmic differentiation and automated decision-making, including injustice by generalisation, treatment of people as mere objects, restrictions on the free development of personality and informational self-determination, accumulation effects and growing inequality as well as risks to societal goals of equality or social policy. In these cases, there is a need for reforms of the anti-discrimination and data protection law, but also for societal considerations and definitions of which kinds of algorithmic differentiations are considered acceptable in a society in order to protect fundamental rights and values. Last but not least, the study discusses tasks for anti-discrimination agencies and equality bodies, ranging from the identification and proof of algorithm-based discrimination to preventive and cooperative actions.

/pdf/risks-of-discrimination-through-the-use-of-algorithms-a-44194s3co9.pdf

Risks of Discrimination through the Use of Algorithms. A study compiled with a grant from the Federal Anti-Discrimination Agency

Organizations are increasingly employing complex black-box machine learning models in high-stakes decision-making. A popular approach to addressing the problem of opacity of black-box machine learning models is the use of post-hoc explainability methods. These methods approximate the logic of underlying machine learning models with the aim of explaining their internal workings, so that human examiners can understand them. In turn, it has been alluded that the insights from post-hoc explainability methods can be used to help regulate black-box machine learning. This article examines the validity of these claims. By examining whether the insights derived from post-hoc explainability methods in post-model deployment can prima facie meet legal definitions in European (read European Union) non-discrimination law, we argue that machine learning post-hoc explanation methods cannot guarantee the insights they generate. Ultimately, we argue that the use of post-hoc explanatory methods is useful in many cases, but that these methods have limitations that prohibit reliance as the sole mechanism to guarantee fairness of model outcomes in high-stakes decision-making. By way of an ancillary function, the inadequacy of European Non-Discrimination Law for algorithmic decision-making is demonstrated too. 

https://link.springer.com/content/pdf/10.1007/s43681-022-00142-y.pdf

Explainable artificial intelligence (XAI) post-hoc explainability methods: risks and limitations in non-discrimination law

Regulierung intelligenter Systeme

Regulating Artificial Intelligence

/pdf/unconstitutional-constitutional-amendments-4x5educzyv.pdf

Unconstitutional Constitutional Amendments

Nationale Identität und Verfassungsidentität. Schutzgehalte, Instrumente, Perspektiven

For a long time, EU institutions have emphasized the connection between one of the most important concepts of the integration method, mutual recognition, and the presence of mutual trust between EU Member States. Only recently, the ECJ reaffirmed in its Opinion 2/13 that mutual trust is at the heart of the EU and a “fundamental premiss” of the European legal structure. But can law really restore, advance or even govern by trust? This question is crucial for the EU of today, which finds itself in the midst of a severe crisis of trust. For the EU as a community “based on the rule of law” generating trust through law might seem the natural, maybe the only politically viable response to a crisis of trust. Nevertheless, even if one agrees that the rule of law requires people to place trust in legal rules, and that courts and administrative agencies need to trust each other in order to work efficiently and consistently, how would legal rules be able to generate or promote trust? Moreover, isn’t it deeply rooted in our ideas about constitutional government that democratic law must institutionalize mutual distrust rather than govern by trust? These conceptual and normative objections did not stop the European Union from pursuing the project of trust-building through law in one of the most sensitive areas of EU law, judicial cooperation in civil and criminal matters. This Article will ask whether the project to promote trust through law is a promising one, and, eventually, how to reinterpret statutory provisions and legal principles that purport to generate trust amongst their addressees.

Thomas Wischmeyer

Papers

Regulierung intelligenter Systeme

Regulating Artificial Intelligence

Unconstitutional Constitutional Amendments

Nationale Identität und Verfassungsidentität. Schutzgehalte, Instrumente, Perspektiven

Generating Trust Through Law? – Judicial Cooperation in the European Union and the 'Principle of Mutual Trust'