Search

Det här kapitlet pekar på hur europeisk AI-reglering brottas med att reglera ett så exceptionellt rörligt mål som AI-utvecklingen, och visar hur dessa insikter är relevanta för svensk offentlig förvaltning och dess tillsynsmyndigheter när den europeiska AI-förordningen blir tillämplig. Kapitlet pekar på behov av strukturer för kompetensutveckling, samverkan och aktiv tolkning och kommunikation av både redan befintlig lagstiftning såväl som av kommande tillämpning av AI-förordningen. Särskild uppmärksamhet riktas mot generativ AI eftersom det är ett särskilt rörligt fält som påtagligt påverkat lagstiftningsprocessen.

Extant literature points to how the risk of discrimination is intrinsic to AI systems owing to the dependence on training data and the difficulty of post hoc algorithmic auditing. Transparency and auditability limitations are problematic both for companies’ prevention efforts and for government oversight, both in terms of how artificial intelligence (AI) systems function and how large‐scale digital platforms support recruitment processes. This article explores the risks and users’ understandings of discrimination when using AI and automated decision‐making (ADM) in worker recruitment. We rely on data in the form of 110 completed questionnaires with representatives from 10 of the 50 largest recruitment agencies in Sweden and representatives from 100 Swedish companies with more than 100 employees (“major employers”). In this study, we made use of an open definition of AI to accommodate differences in knowledge and opinion around how AI and ADM are understood by the respondents. The study shows a significant difference between direct and indirect AI and ADM use, which has implications for recruiters’ awareness of the potential for bias or discrimination in recruitment. All of those surveyed made use of large digital platforms like Facebook and LinkedIn for their recruitment, leading to concerns around transparency and accountability—not least because most respondents did not explicitly consider this to be AI or ADM use. We discuss the implications of direct and indirect use in recruitment in Sweden, primarily in terms of transparency and the allocation of accountability for bias and discrimination during recruitment processes.

Extant literature points to how the risk of discrimination is intrinsic to AI systems owing to the dependence on training data and the difficulty of post hoc algorithmic auditing. Transparency and auditability limitations are problematic both for companies' prevention efforts and for government oversight, both in terms of how artificial intelligence (AI) systems function and how large-scale digital platforms support recruitment processes. This article explores the risks and users' understandings of discrimination when using AI and automated decision-making (ADM) in worker recruitment. We rely on data in the form of 110 completed questionnaires with representatives from 10 of the 50 largest recruitment agencies in Sweden and representatives from 100 Swedish companies with more than 100 employees ("major employers"). In this study, we made use of an open definition of AI to accommodate differences in knowledge and opinion around how AI and ADM are understood by the respondents. The study shows a significant difference between direct and indirect AI and ADM use, which has implications for recruiters' awareness of the potential for bias or discrimination in recruitment. All of those surveyed made use of large digital platforms like Facebook and LinkedIn for their recruitment, leading to concerns around transparency and accountability-not least because most respondents did not explicitly consider this to be AI or ADM use. We discuss the implications of direct and indirect use in recruitment in Sweden, primarily in terms of transparency and the allocation of accountability for bias and discrimination during recruitment processes.

By introducing the concept of high-reach AI, this paper focuses on AI systems whose widespread use may generate significant risks for both individuals and societies. While some of those risks have been recognised under the AI Act, we analyse the rules laid down by the Digital Services Act (DSA) for recommender systems used by dominant social media platforms as a prominent example of high-reach AI. Specifically, we examine transparency provisions aimed at addressing adverse effects of these AI technologies employed by social media very large online platforms (VLOPs). Drawing from AI transparency literature, we analyse DSA transparency measures through the conceptual lens of horizontal and vertical transparency. Our analysis indicates that while the DSA incorporates transparency provisions in both dimensions, the most progressive amendments emerge within the vertical transparency, for instance, by the introduction of the systemic risk assessment mechanism. However, we argue that the true impact of the new transparency provisions extends beyond their mere existence, emphasising the critical role of oversight entities in implementation and application of the DSA. Overall, this study highlights the paramount importance of vertical transparency in providing a comprehensive understanding of the aggregated risks associated with high-reach AI technologies, exemplified by social media recommender systems., QC 20240709

Modelling railway projects has a main challenge in the discrete element method (DEM). The granular material of the embankment consists of millions of fine angular particles which are difficult to model due to the long computational time. The long computational time also prevents the modeling of the higher number of loading cycles. As a result, researchers prefer to simulate the project in 2D to accelerate the simulation. While 2D simulations present a seemingly simple option for modeling railways, they tend to oversimplify the intricacies of particle interactions and the distribution of stress. Nonetheless, the extent to which these simplifications affect the authenticity of the simulations has remained ambiguous. In this study, the periodic cell replication method is used to build extensive long railway tracks significantly faster than conventional methods. Then, this DEM model is calibrated against the measurement results of a physical full-scale ballasted track. The model is then used to simulate several railway projects with different initial particle arrangements and model dimensions in both 2D and 3D. The results show that the 2D models are more dependant on the initial particle arrangement which shows different behavior for the same model. In addition, 2D simulations are incapable of reproducing the principal stress rotation in granular layers due to the moving load of the train wheel. As a result, 3D DEM simulations using the periodic cell replication method is suggested for studying the railway tracks., QC 20240527

In the context of geo-infrastructures and specifically tunneling projects, analyzing the large-scale sensor-based measurement-while-drilling (MWD) data plays a pivotal role in assessing rock engineering conditions. However, handling the big MWD data due to multiform stacking is a time-consuming and challenging task. Extracting valuable insights and improving the accuracy of geoengineering interpretations from MWD data necessitates a combination of domain expertise and data science skills in an iterative process. To address these challenges and efficiently normalize and filter out noisy data, an automated processing approach integrating the stepwise technique, mode, and percentile gate bands for both single and peer group-based holes was developed. Subsequently, the mathematical concept of a novel normalizing index for classifying such big datasets was also presented. The visualized results from different geo-infrastructure datasets in Sweden indicated that outliers and noisy data can more efficiently be eliminated using single hole-based normalizing. Additionally, a relational unified PostgreSQL database was created to store and automatically transfer the processed and raw MWD as well as real time grouting data that offers a cost effective and efficient data extraction tool. The generated database is expected to facilitate in-depth investigations and enable application of the artificial intelligence (AI) techniques to predict rock quality conditions and design appropriate support systems based on MWD data., QC 20240429

Construction and infrastructure projects account for a large share of global emissions. Included in these projects are geotechnical engineering works, which cause environmental impact and costs throughout their lifecycle. In the geotechnical design process, the geotechnical engineer often has design choices where different methods and materials can be used. If assessments of environmental impact and costs were to be integrated into the geotechnical design process, the geotechnical work could be made more sustainable. As a result of this, there is a need for research on integrating the assessments of environmental impact and costs in the geotechnical design process. This paper, therefore, presents a methodology showing how Life Cycle Assessment (LCA) and Life Cycle Cost Analysis (LCCA) could be integrated into the geotechnical design process. The presented methodology is then illustrated in a case study for assessing the climate impact and costs of high-speed railway embankment fill methods., QC 20240612

There is an increasing demand to utilize technological possibilities in the Nordic public sector. Automated decision-making (ADM) has been deployed in some areas towards that end. While ADM is associated with a range of benefits, research shows that its use, with elements of AI, also implicates risks of discrimination and unfair treatment, which has stimulated a flurry of normative guidelines. This article seeks to explore how a sample of these international high-level principled ideas on fairness translate into the specific governance of ADM in national public-sector authorities in Sweden. It does so by answering the question of how ideas about AI ethics and fairness are considered in relation to regulation on anti-discrimination in Swedish public-sector governance. By using a Scandinavian institutionalist approach to translation theory, we trace how ideas about AI governance and public-sector governance translate into state-authority practice; specifically, regarding the definition of ADM, how AI has impacted it as both discourse and technology, and the ideas of ‘ethicsʼ and ‘discriminationʼ. The results indicate that there is a variance in how different organizations understand and translate ideas about AI ethics and discrimination. These tensions need to be addressed in order to develop AI governance practices.

Det är mycket rörelse på AI-fronten nu – både rättsligt, tekniskt och politiskt. Stefan Larsson, jurist och docent i teknik och social förändring vid Lunds universitet, placerar in EU-lagstiftarens AI-förordning i en både global och svensk kontext. Han beskriver processen bakom lagstiftningen, de etiska och principiella grunderna, och hur EU-lagstiftaren hanterar AI-innovationernas explosionsartade rörlighet med flexibla element i AI-förordningen, som på samma gång kan skapa dynamisk styrning och rättslig oförutsägbarhet.