Browsing by Author "Wright, David"

Now showing 1 - 9 of 9
  • Thumbnail Image
    ItemOpen Access
    Artificial intelligence for human flourishing – Beyond principles for machine learning
    (Elsevier, 2021) Stahl, Bernd Carsten, 1968-; Andreou, Andreas; Brey, Philip; Hatzakis, Tally; Kirichenko, Alexey; Macnish, Kevin; Laulhe Shaelou, Stephanie; Patel, Andrew; Ryan, Mark; Wright, David
    The technical and economic benefits of artificial intelligence (AI) are counterbalanced by legal, social and ethical issues. It is challenging to conceptually capture and empirically measure both benefits and downsides. We therefore provide an account of the findings and implications of a multi-dimensional study of AI, comprising 10 case studies, five scenarios, an ethical impact analysis of AI, a human rights analysis of AI and a technical analysis of known and potential threats and vulnerabilities. Based on our findings, we separate AI ethics discourse into three streams: (1) specific issues related to the application of machine learning, (2) social and political questions arising in a digitally enabled society and (3) metaphysical questions about the nature of reality and humanity. Human rights principles and legislation have a key role to play in addressing the ethics of AI. This work helps to steer AI to contribute to human flourishing.
  • Thumbnail Image
    ItemOpen Access
    Assessing the Specificity and Accuracy of Accent Judgments by Lay Listeners
    (Sage, 2022-06-19) Braber, Natalie; Smith, Harriet; Wright, David; Hardy, Alexander; Robson, Jeremy
    Historically, there has been less research carried out on earwitness than eyewitness testimony. However, in some cases, earwitness evidence might play an important role in securing a conviction. This paper focuses on accent which is a central characteristic of voices in a forensic linguistic context. The paper focuses on two experiments (Experiment 1, n = 41; Experiment 2, n = 57) carried out with participants from a wide range of various locations around the United Kingdom to rate the accuracy and confidence in recognizing accents from voices from England, Scotland, Wales, Northern Ireland, and Ireland as well as looking at specificity of answers given and how this varies for these regions. Our findings show that accuracy is variable and that participants are more likely to be accurate when using vaguer descriptions (such as “Scottish”) than being more specific. Furthermore, although participants lack the meta-linguistic ability to describe the features of accents, they are able to name particular words and pronunciations which helped them make their decision.
  • No Thumbnail Available
    ItemOpen Access
    Ethics and Privacy in AI and Big Data: Implementing Responsible Research and Innovation
    (IEEE, 2018-06-25) Stahl, Bernd Carsten, 1968-; Wright, David
    Emerging combinations of artificial intelligence, big data and the applications these enable are receiving significant media and policy attention. Much of the attention concerns privacy and other ethical issues. In our paper, we suggest that what is needed now is a way to comprehensively understand these issues and find mechanisms of addressing them that involve stakeholders, including civil society, to ensure that these technologies’ benefits outweigh their disadvantages. We suggest that the concept of responsible research and innovation (RRI) can provide the framing required to act with a view to ensuring that the technologies are socially acceptable, desirable and sustainable. We draw from our work on the Human Brain Project, one potential driver for the next generation of these technologies, to discuss how RRI can be put in practice.
  • Thumbnail Image
    ItemOpen Access
    Evaluating earwitness identification procedures: adapting pre-parade instructions and parade procedure
    (Taylor and Francis, 2022-10-06) Smith, Harriet; Roeser, Jens; Pautz, Nikolas; Davis, Josh P; Robson, Jeremy; Wright, David; Braber, Natalie; Stacey, Paula C.
    Voice identification parades can be unreliable, as earwitness responses are error-prone. In this paper we tested performance across serial and sequential procedures, and varied pre-parade instructions, with the aim of reducing errors. The participants heard a target voice and later attempted to identify it from a parade. In Experiment 1 they were either warned that the target may or may not be present (standard warning) or encouraged to consider responding “not present” because of the associated risk of a wrongful conviction (strong warning). Strong warnings prompted a conservative criterion shift, with participants less likely to make a positive identification regardless of whether the target was present. In contrast to previous findings, we found no statistically reliable difference in accuracy between serial and sequential parades. Experiment 2 ruled out a potential confound in Experiment 1. Taken together, our results suggest that adapting pre-parade instructions provides a simple way of reducing the risk of false identifications
  • Thumbnail Image
    ItemOpen Access
    Exploring ethics and human rights in artificial intelligence - A Delphi study
    (Elsevier, 2023-03-23) Stahl, Bernd; Brooks, Laurence; Hatzakis, Tally; Santiago, Nicole; Wright, David
    Ethical and human rights issues of artificial intelligence (AI) are a prominent topic of research and innovation policy as well as societal and scientific debate. It is broadly recognised that AI-related technologies have properties that can give rise to ethical and human rights concerns, such as privacy, bias and discrimination, safety and security, economic distribution, political participation or the changing nature of warfare. Numerous ways of addressing these issues have been suggested. In light of the complexity of this discussion, we undertook a Delphi study with experts in the field to determine the most pressing issues and prioritise appropriate mitigation strategies. The results of the study demonstrate the difficulty of defining clear priorities. Our findings suggest that the debate around ethics and human rights of AI would benefit from being reframed and more strongly emphasising the systems nature of AI ecosystems.
  • Thumbnail Image
    ItemOpen Access
    Policy scenarios as an instrument for policymakers
    (Elsevier, 2020) Wright, David; Stahl, Bernd Carsten, 1968-; Hatzakis, Tally
    Scenarios are a methodology of futures and foresight research that has been established for more than half a century. Despite the rich literature on scenarios and ways of constructing them, the current scenario methodologies, current approaches to scenario constructions are often not well aligned with the needs of policymakers. Emerging technologies that are likely to achieve high social relevance in the short to mid-term future and that call for quick policy interventions are difficult to reflect using current scenario approaches. We have therefore developed a new type of scenario that we call a policy scenario. This paper develops the principles and justification for policy scenarios. We provide a detailed description of how they can be constructed, focusing on their key characteristics of policy requirements, plausibility, probability, credibility, expertise, objectivity and legitimacy. Following our stakeholder-based approach allows researchers to construct scenarios that are uniquely suited to inform policymakers and, in effect, the policy development process.
  • No Thumbnail Available
    ItemOpen Access
    The pragmatic functions of ‘respect’ in lawyers’ courtroom discourse: a case study of Brexit hearings
    (2021-11-17) Robson, Jeremy; Murray - Edwards, Helen; Braber, Natalie; Wright, David
    This paper is a corpus-assisted discourse analysis of the use of the word respect by the main advocates in the High Court and Supreme Court hearings of R v Secretary of State for Exiting the European Union (the ‘Brexit case’). Courtroom discourse has received substantial research attention in pragmatics, and previous work has largely focused on notions of face and im/politeness exhibited in power-asymmetric encounters between lawyers and witnesses in hostile cross-examination. In contrast, this paper focuses on lawyer-lawyer and lawyer-judge interaction in appellate hearings and explores the ways in which advocates negotiate the task of making face-threats that are inherent to the discourse situation, while maintaining the levels of professional courtesy demanded by the institution. The word respect has a particular role in managing this balance, and has attached to it well-established implicit, indexical and professional meanings within the judiciary. The corpus analysis here shows that, although the advocates in question use respect in seemingly formulaic and ritualised ways, it is used to achieve multiple facework and interactional goals. Throughout the analysis we see advocates use respect when (dis)agreeing with judges, challenging opposing counsel and making recommendations to the court.
  • Thumbnail Image
    ItemOpen Access
    A systematic review of artifcial intelligence impact assessments
    (Springer, 2023-03-24) Stahl, Bernd; Antoniou, Josephina; Bhalla, Nitika; Brooks, Laurence; Jansen, Philip; Lindqvist, Blerta; Kirichenko, Alexey; Marchal, Samuel; Rodrigues, Rowena; Warso, Zuzanna; Wright, David; Santiago, Nicole
    Artificial intelligence (AI) is producing highly beneficial impacts in many domains, from transport to healthcare, from energy distribution to marketing, but it also raises concerns about undesirable ethical and social consequences. AI impact assessments (AI-IAs) are a way of identifying positive and negative impacts early on to safeguard AI’s benefits and avoid its downsides. This article describes the first systematic review of these AI-IAs. Working with a population of 181 documents, the authors identified 38 actual AI-IAs and subjected them to a rigorous qualitative analysis with regard to their purpose, scope, organisational context, expected issues, timeframe, process and methods, transparency and challenges. The review demonstrates some convergence between AI-IAs. It also shows that the field is not yet at the point of full agreement on content, structure and implementation. The article suggests that AI-IAs are best understood as means to stimulate reflection and discussion concerning the social and ethical consequences of AI ecosystems. Based on the analysis of existing AI-IAs, the authors describe a baseline process of implementing AI-IAs that can be implemented by AI developers and vendors and that can be used as a critical yardstick by regulators and external observers to evaluate organisations’ approaches to AI.
  • No Thumbnail Available
    ItemOpen Access
    Voice parade procedures: Optimising witness performance
    (Taylor & Francis, 2019-10-08) Robson, Jeremy; Smith, Harriet; Braber, Natalie; Wright, David; Roeser, Jens; Stacey, Paula; Bird, Katherine
    Unfamiliar voice identification is error-prone. Whilst the investigation of system variables may indicate ways of boosting earwitness performance, this is an under-researched area. Two experiments were conducted to investigate how methods of presenting voices during a parade affect accuracy and self-rated confidence. In each experiment participants listened to a target voice, and were later asked to identify that voice from a nine-person target present or target absent parade. In Experiment 1, accuracy did not vary across parades comprising 15 or 30 s sample durations. Overall, when the target was present, participants correctly identified the target voice with 39% accuracy. However, when the target was absent, participants correctly rejected the parade 6% of the time. There was no relationship between accuracy and confidence. In Experiment 2, performance with a serial procedure, in which participants responded after hearing all nine voices, was compared with a sequential procedure, in which participants made a decision after listening to each voice. Overall accuracy was higher with the sequential procedure. These results highlight the importance of system variable research in voice identification. Different methods of presenting voices have the potential to support higher levels of accuracy than the procedure currently recommended in England and Wales.