52 research outputs found
Schools of AI in the Public Sector: Fairness and Accountability Concerns
No full textAs decision-making algorithms become more prevalent in society, the importance efficiency and problem-solving abilities come into question when predictions impact individuals’ lives. High-risk applications require trusted AI systems to be designed with fairness and accountability; such trust and consideration is essential for public acceptance and successful deployment.
Despite growing advocacy for ethical and trustworthy AI, along with the emergence of guidelines like the EU AI Act, controversies surrounding AI persist in the media. Public sector AI systems are being implemented haphazardly, whether in judicial decision-making, healthcare diagnostics, or social welfare distribution. These high-risk applications directly affect citizens’ quality of life, highlighting the need for a critical assessment of how AI are being designed and deployed in the public sector.
My thesis explores the integration of fairness, accountability, and uncertainty in public sector AI to assess whether these systems are appropriately designed, effectively adapted, and capable of enhancing societal well-being. The research aims to provide actionable insights for designing AI systems that align with public sector needs and maximize societal benefits
AI for public good: reorientating fairness, accountability and uncertainty in public sector governance
No full textDeploying AI in uncertain environments: a technical limitation or a human characteristic?
No full textThe relationship between AI and uncertainty in high-stakes public environments has not yet been given the attention that it requires. While technical literature often frames uncertainty as a limitation that should be resolved or minimised, this project draws attention to an alternative interpretation: uncertainty as a fundamental and valuable component of human judgment, particularly within many aspects of public sector decision-making, and therefore minimising uncertainty to design more effective AI can become undesirable. My research investigates how AI systems designed for predictability, consistency, and optimization struggle to operate effectively in environments where discretion, ambiguity, and pluralism are not only unavoidable but often necessary. This project advances the conceptual understanding of uncertainty in AI ethics and governance while also offering early empirical insights through experiments with large language models in legal interpretive tasks. The overarching aim is to develop normative and technical guidance for building AI systems that align more meaningfully with the social and institutional functions of uncertainty. Additionally, I acknowledge the benefits of meaningfully minimising environmental uncertainties for AI systems and my future work aspires to produce a framework to help guide when adaptations to reduce uncertainty for public sector AI are permittable and when they should not be made to ensure the inherent humanness of society remains intact
We built the wrong AI for the right problems: why public sector AI gets it wrong from the start
No full textThe highly dynamic heterochromatin protein Swi6 mediates degradation of heterochromatic transcripts
Get PDFThe aim of my thesis was to investigate the mechanism of heterochromatin repression mediated by heterochromatin protein Swi6 in S. pombe.
Research over the years challenged the view of heterochromatin as a static and transcriptionally inert structure. Especially in fission yeast it has become clear that heterochromatin silencing requires not only the action of chromatin modifying factors, but also transcriptional activity and RNA degradation processes. Moreover, heterochromatin protein Swi6 was shown to be highly dynamic, unlike what was expected for a protein that is perceived as a major structural component of heterochromatin. Yet, the prevailing model of heterochromatin establishment and spreading is thought to occur by iterative HP1 binding to methylated H3K9 and recruitment of histone methylation activity.
Driven by recent findings in our lab that described a new role for Swi6 in repression and provided a possible explanation for its dynamic behavior, I set out to investigate the mechanism in vivo by studying Swi6 dynamics. Therefore, a major focus of my PhD was to establish a suitable, robust microscopy-based method that allowed me to follow rapid dynamics and produce reliable data. The work I have done challenged the role of Swi6 in heterochromatin maintenance and spreading, but coincides with a clear involvement in sustaining tight repression. While H3K9me levels remained high in the absence of Swi6 and even spread into neighboring regions, heterochromatic transcript levels increased. These observations revealed unanticipated functions for Swi6 and made us reconsider the mechanism of Swi6-mediated silencing. As previously proposed, Swi6 could function as a co-transcriptional checkpoint that mediates RNA degradation (Keller et al., 2012). In this model RNA binds to Swi6 and gets primed for destruction as it is handed over to the RNA decay machinery, involving Cid14 and the exosome or the RNAi machinery. The target specificity depends on the epigenetic make-up of the locus, meaning the recognition of H3K9me by the CD of Swi6, while RNA binding occurs in a sequence independent manner (Keller et al., 2012). Therefore, additional processes that confer specificity, like siRNAs, are needed to ensure correct targeting of H3K9me marks to trigger Swi6-mediated turnover of unwanted RNA transcripts and not of any other random region in the genome
Mapping dominant AI schools to multiple accountability types
No full textPurpose – As AI algorithms become central to public policy development and delivery, ensuring accountability in automated public services is crucial. This paper extends algorithmic accountability research by proposing a framework to help AI designers and public practitioners understand AI’s impact on diverse accountability relationships and identifies how AI systems may be better designed for greater public benefit.Design/Methodology/Approach – The study employs an inductive approach, combining established frameworks from accountability studies, computer science, and public governance. By evaluating the conceptual and technical characteristics of the two most dominant AI paradigms (connectionist and symbolic), this study systematically maps their compatibility with four formal accountability forums across three phases of accountability. The resulting conceptual mapping framework highlights the trade-offs and alignment of AI design choices with diverse public accountability demands.Findings – Findings indicate that a singular AI paradigm cannot simultaneously provide effective accountability to multiple forums. Current public AI deployment practices appear to prioritise internal technocratic objectives over designing algorithmic systems towards effective transparent accountability processes, raising concerns about alignment with public accountability standards.Practical Implications – The proposed mapping framework provides a practical tool for public practitioners and AI system designers, offering insights into how AI systems might be tailored to enhance public sector accountability relationships.Originality – To the authors' best knowledge, this study is the first to directly explore the compatibility of AI paradigms with different accountability requirements, offering a novel perspective on aligning AI design with effective multi-forum accountability
Evolving generative AI: entangling the accountability relationship
No full textSince ChatGPT’s debut, generative AI technologies have surged in popularity within the AI community. Recognized for their cutting- edge language processing capabilities, these excel in generating human-like conversations, enabling open-ended dialogues with end-users. We consider that the future adoption of generative AI for critical public domain applications transforms the accountability relationship. Previously characterized by the relationship between an actor and a forum, the introduction of generative systems complicates accountability dynamics as the initial interaction shifts from the actor to an advanced generative system. We conceptualise a dual-phase accountability relationship involving the actor, the forum and the generative AI as a foundational approach to understanding public sector accountability in the context of these technologies. Focusing on integrating generative AI for assisting healthcare triaging, we identify potential challenges introduced for maintaining effective accountability relationships, highlighting concerns that these technologies relegate actors to a secondary phase of accountability and creates a disconnect between government actors and citizens. We suggest recommendations aimed at disentangling the complexities generative systems bring to the accountability relationship. As we speculate on the technologies disruptive impact on accountability, we urge public servants, policymakers, and system designers to deliberate on the potential accountability impact generative systems produce prior to their deployment
On the link between forward energy prices: A nonlinear panel cointegration approach
Get PDFThis paper investigates the relationship between forward prices of oil, gas, coal, and electricity using a nonlinear panel cointegration framework. To this end, we consider a panel of 35 maturities and control for the economic and financial environment using equity futures prices. Estimating the cointegrating relationship, we find that oil, gas and coal forward prices are positively linked, while the negative link between oil and electricity prices is consistent with a substitution effect between the two energy sources on the long run. Estimating panel smooth transition regression (PSTR) models, we show that the forward oil price adjustment process toward its equilibrium value is nonlinear and asymmetric, putting forward the key role played by self-sustaining dynamics and speculation phenomena.forward energy prices, speculation, panel cointegration, nonlinear model, PSTR
Dissecting the PPP Puzzle: The Unconventional Roles of Nominal Exchange Rate and Price Adjustment
Get PDFThe conventional view, as expounded by sticky-price models, is that price adjustment determines the PPP reversion rate. This study examines the mechanism by which PPP deviations are corrected. Nominal exchange rate adjustment, not price adjustment, is shown to be the key engine governing the speed of PPP convergence. Moreover, nominal exchange rates are found to converge much more slowly than prices. With the reversion being driven primarily by nominal exchange rates, real exchange rates also revert at a slower rate than prices, as identified by the PPP puzzle (Rogoff, 1996).
Time Series Analysis
Get PDFWe provide a concise overview of time series analysis in the time and frequency domains, with lots of references for further reading.time series analysis, time domain, frequency domain
- …
