https://www.mdu.se/

mdu.sePublications
Change search
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Large Language Models and Self-determination: In search for a framework of digitally humane Artificial Intelligence
Mälardalen University, School of Innovation, Design and Engineering.
2025 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

Exploring the potential integration of Digital Humanism within Digital SocialInnovations (DSIs), this research focuses on the responsible application of LargeLanguage Models (LLMs) to enhance user autonomy and self-determination. WhileDigital Humanism emphasizes protecting human values and autonomy in digitaltechnology design, a paradox exists where systems intended to enhance capabilitiescan inadvertently undermine human agency. This research investigates how tosupport human capacities through the deployment of LLMs. In this thesis, themethodological approach employed integrates the multidisciplinary orientation ofsocial innovation research and argumentative principles rooted in pragmaticjustification through collective discourse, drawing inspiration from Rorty'sphilosophy (cited in Philosophy Overdose; 2023). Design research provides theprimary methodological structure, with interactive research serving as the basis forquality control. This strategy begins with reviewing existing literature to establishthe problem and potential solution space. Subsequently, the proposed innovationundergoes a process of iterative refinement and construction, informed by collectivecriticism and building, ultimately seeking validation through its demonstratedutility through the emergent consensus within the relevant community. BehavioralEconomics insights were incorporated to analyze cognitive biases and theirinfluence on interactions with AI systems. Insights gained from interpretation andsynthesis revealed a tension between the streamlined utility often prioritized indigital platforms and the value of diverse perspectives. Challenges identifiedinclude the potential for current AI development, especially concerning LLMs, tocreate innovation barriers due to dominant actors and proprietary models. Concernsalso arose regarding the difficulty in ensuring the reliability of LLMs and possibleconsequences. With the purpose of exploring, understanding, and proposing adefensible solution, a path forward is presented involving decentralized,community-governed AI systems, inspired by Elinor Ostrom's work on commonpool resources. Such system is designed to integrate localized values andknowledge, and ethically advance behavioral insights to promote well-being andinformed decision-making. This approach aims to counteract homogenization,emphasizing the fundamental contribution of continuous, diverse human-authoredcontent for maintaining epistemic vitality. While outlining this conceptual approach, the research acknowledges remaining questions regarding practical implementationand scaling.

Place, publisher, year, edition, pages
2025.
National Category
Other Social Sciences
Identifiers
URN: urn:nbn:se:mdh:diva-73099OAI: oai:DiVA.org:mdh-73099DiVA, id: diva2:1992046
Subject / course
Innovation & design
Supervisors
Examiners
Available from: 2025-09-09 Created: 2025-08-26 Last updated: 2025-10-10Bibliographically approved

Open Access in DiVA

No full text in DiVA

Search in DiVA

By author/editor
Carlos Andrade Uzeda Accioly, Ana Gabriela
By organisation
School of Innovation, Design and Engineering
Other Social Sciences

Search outside of DiVA

GoogleGoogle Scholar

urn-nbn

Altmetric score

urn-nbn
Total: 76 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf