Due to recent advances within the field of machine learning and computing power becoming more readily available, the use of machine learning within the field of psychology has increased. However,...Show moreDue to recent advances within the field of machine learning and computing power becoming more readily available, the use of machine learning within the field of psychology has increased. However, potential remains for greater use of machine learning within the field of psychology. In this study the usability and performance of 3 machine learning models namely K-nearest neighbors, the Random forest, and the Support vector machine algorithms were assessed when predicting gender, marital status, and family size from Big 5 personality measures and the Holland Code Career Test. Repeated cross-validation was combined with grid search to ensure performance measure accuracy and to optimize model accuracy and F1-score. The performances of the 3 models were compared to the performance of logistic regression to assess whether these models could outperform a model regularly used within psychology. The 3 models consistently outperformed the logistic regression under almost all conditions and proved far superior for groups sizes over 500 even outperforming logistic regression by 10 percentage points under some conditions. However, caution was advised due to wide confidence intervals for small group sizes (n ≤ 200). Therefore, a study was proposed with the aim to enhance predictions for small group sizes, focusing on feedforward neural networks, known to be able to capture complex relationships even with limited data. Addressing these aspects could improve the usability of machine learning in psychology settings involving small group sizes.Show less
In recent years, Automated Influence, understood as “the use of artificial intelligence to collect, integrate and analyse people’s data, and to deliver targeted interventions based on this analysis...Show moreIn recent years, Automated Influence, understood as “the use of artificial intelligence to collect, integrate and analyse people’s data, and to deliver targeted interventions based on this analysis, intended to shape their behaviour” (familiarly referred to as ‘algorithms’) has stirred up many debates among the public, as well as within academia (Benn & Lazar 2022, 127). While much of the discussion has focused primarily on issues of privacy in the light of Big Data, this thesis seeks to analyze how Automated Influence impacts the deliberative, discursive, and fundamentally social space on which society depends on, in particular for collective decision-making/politics. I argue that Automated Influence deployed on social media platforms violates people’s fundamental interest in social agency, which is defined as the ability of a person to act and reflect on her own motives all the while taking part in the fundamentally social process of forming, defending, and adapting the reasons according to which she acts. Moreover, it undermines people’s autonomy and social trust, which both serve as preconditions for their exercise social agency. After reviewing contemporary EU regulation seeking to address some of the problematic aspects related to Automated Influence, I explain why there cannot be a purely top-down approach to mitigating the harms emanating from Automated Influence, which results in my conclusion that only through educating people about its potential harms could mitigate the problem in the long run.Show less
The (hypothetical) deployment of Lethal Autonomous Weapons (LAWs) challenges the way in which we conceptualize moral responsibility. The emergence of LAWs have added an autonomously acting non...Show moreThe (hypothetical) deployment of Lethal Autonomous Weapons (LAWs) challenges the way in which we conceptualize moral responsibility. The emergence of LAWs have added an autonomously acting non-human entity to a moral responsibility framework which is inextricably linked to human nature and moral capacity, which LAWs neither have nor possess. This leaves open a responsibility gap in which it becomes unclear who exactly is responsible for the outcome of the decisions made by LAWs. Although several solutions have been proposed to solve the gap, such as the concept of meaningful control or role-specific responsibility, I find that they cannot sufficiently address the responsibility gap. The concept of meaningful human control is inadequate for the complex and chaotic environment of warfare, particularly when introducing powerful weapons that push the boundaries of human capability. While role-responsibility considers the collective nature of the military and the entire chain of command, it faces challenges in accounting for the problem of many hands and the emergent behavior of autonomous weapons that cannot be directly attributed to a specific part of the system or individual. Especially in a value-loaded and ethically charged environment such as war, where choices regarding life or death are a routine matter, there is no room for obscured responsibility. Without proper responsibility, one cannot justify the introduction of LAWs onto the battlefield.Show less
Bachelor thesis | Liberal Arts and Sciences: Global Challenges (LUC) (BA/BSc)
closed access
Aims: The aim of this research is to understand how effective contraceptive devices using algorithms based on basal body temperature can be at preventing unwanted pregnancies. Methods: Three...Show moreAims: The aim of this research is to understand how effective contraceptive devices using algorithms based on basal body temperature can be at preventing unwanted pregnancies. Methods: Three different databases (PubMed, Embase and Web of Science) were searched based on specific keyword profiles to yield a maximum of relevant articles. A total of 85 citations were reviewed from which 10 were included in this study as they met the precise inclusion criteria. The snowballing method was then employed to retrieve an additional 6 relevant articles. These studies were then rated (weak, moderate or strong) according to four criteria of quality assessment. Results: A total of 6 different devices were reviewed in this study: Natural Cycles, Daysy, LadyComp, Pearly, Cyclotest 2 Plus and Bioself 110 (and 2000). All of the 16 studies included in this review supported the idea that devices using algorithms based on basal body temperature were effective at preventing unwanted pregnancy. Considering all methods, the percentage of fertile days wrongly identified as fertile ranged between 0.07 – 7.5 (%) with some methods being more effective than others. The typical use Pearl index ranged from 0.5 to 9.8 and perfect use Pearl index from 0.7 to 2.0. Conclusion: The studies included in this systematic review demonstrate that algorithms based on basal body temperature can be effective birth control methods. The six different methods were all considered effective but difference in the quality of the studies included for each method might compromise some of these results.Show less
This work critically assesses the idea of replacing political representation based on elections and politicians by big data-driven algorithms. The rapid digitalization and datafication of our world...Show moreThis work critically assesses the idea of replacing political representation based on elections and politicians by big data-driven algorithms. The rapid digitalization and datafication of our world is fuelling the debate on democratic theory. Can the potential of new ICTs be harnessed to work for the benefit of democracy? Using Gijs van Oenen’s account as base, I make the idea of algorithmic representation more concrete by introducing the concept of a Pocket Politician, and by exploring three scenarios of what this could look like. To further unpack this concept I apply two conceptual tools of the ‘constructivist turn’ in democratic theory: Fossen’s logical distinction between dyadic-triadic and Saward’s theory on the representative claim. By doing so, I show that such a new ‘algorithmic’ system of representation would go accompanied with the loss of human intentionality and the loss of visibility. (1) Algorithmic representation blurs the characterization of the citizens that are represented. It is no longer clear as what citizens are represented due to the non-human intentionality of algorithms. And (2) the performance of representation – the representative claims – will not be visible to the citizens, making it a non-transparent form of politics. Combined, these two concessions would restrict citizens’ ability to see and experience politics on both a sensory and mental level. I come back to the three scenarios of a Pocket Politician to discuss whether this is a problem. I argue that applying algorithmic representation could be beneficial for a democracy depending on its particular state and particular needs.Show less
