The Future of Finance and AI in Investment Management

Sammanfattning

Artificial Intelligence (AI) is fundamentally reshaping the landscape of investment management, promising advancements in efficiency, predictive accuracy, and decision optimization. Despite this transformative potential, its widespread adoption faces significant, interconnected challenges concerning algorithmic bias, ethical implications, and the lagging pace of regulatory frameworks. A notable gap exists in empirical research systematically evaluating AI's real-world effectiveness and the impact of these socio-technical barriers through a multi-metric approach. This thesis empirically evaluates the perceived impact of AI on investment management, assessing its efficacy compared to traditional methods while investigating the influence of algorithmic biases, ethical concerns, and regulatory barriers on adoption. Employing a positivist research philosophy and a quantitative, survey-based design, the study utilizes the Goal-Question-Metric (GQM) model to systematically translate the research hypotheses into measurable variables and metrics. Data was collected via a validated, structured questionnaire from 27 AI and finance professionals in the Nordic region, selected through purposive sampling. Analysis incorporated descriptive statistics, measures of central tendency and dispersion, and GQM-aligned quantitative indexing, including the Bias Identification Metric (BIM), Ethical Inhibition Index (EII), and Regulatory Burden Index (RBI). The findings strongly support the transformative potential of AI in enhancing market efficiency, optimization, and decision accuracy compared to traditional methods, with respondents demonstrating overwhelming consensus (>85% agreement) on these benefits (confirming Hypothesis 1). Market efficiency, in this context, refers to the degree to which asset prices fully reflect all available information, ensuring that market participants make optimal decisions based on accurate and timely data. However, the study critically highlights the significant barriers to adoption. A high awareness of algorithmic bias was evident, with >85% of respondents agreeing that data and model biases threaten fairness and outcomes (confirming Hypothesis 2). Ethical concerns, particularly related to transparency and explainability, were perceived as major constraints by >77% of participants, often viewed as more limiting than technical hurdles (confirming Hypothesis 3). Furthermore, regulatory challenges stemming from compliance complexity and jurisdictional variance were cited as substantial inhibitors by >70% of respondents, significantly hindering widespread AI adoption (confirming Hypothesis 4). While AI technologies offer profound strategic and operational advantages in investment management, their responsible and effective integration hinges on proactively addressing algorithmic bias, enhancing ethical governance and transparency, and developing coherent, adaptive regulatory frameworks. The study provides empirical validation for key theoretical concerns discussed in the literature and offers actionable insights for stakeholders navigating the complex evolution of AI in the financial services industry.