Work

Tilannetietoisuuden tukeminen ihmisen ja tekoälyn tiimeissä läpinäkyvyyden ja selitettävyyden avulla

Year: 2026
Type: dissertation
Abstract: This thesis explores the link between situation awareness (SA) and human-AI teaming. The decision-making processes of AI agents can often remain unclear to the users, which negatively affects trust and assessment of an agent’s decisions, thus decreasing efficiency. To increase efficiency, the user’s SA of an agent’s decision-making process can be assessed. Situation awareness can be defined as the user’s ability to perceive, comprehend and project the near future state of the elements in their environment. Displaying information according to these levels of SA can clarify the process behind an AI agent’s decision-making process, thus increasing efficiency. The aim of this thesis was to link transparency and explainability to the three levels of SA in order give suggestions about what information should be displayed to the user about an AI agents decision-making process. A literature review was conducted to articles and studies considering human-AI teaming and SA. The thesis connected findings from these two fields to construct a clear framework. The findings indicate that transparency and explainability are needed in communicating AI’s decision-making processes to users. They also show that by using these features, the user’s SA can be supported through all three levels of SA. For supporting level 1 SA, inputs, outputs, parameters, and active goals should be displayed. For level 2 support, path to current state, causalities and parameters excluded should be shown. For level 3, information displayed should cover intervention needs, limits and effects of human action. In conclusion, this thesis shows that supporting SA in human-AI teams can be done by considering transparency and explainability. Based on the findings, domain specific instructions about information needs could be constructed for fields such as health care, safety, programming and automated vehicle development.
Institution: Aalto University
Language: English
Open Access status: green
OpenAlex ID: W7221231618
Cites: 22
Cited by:
Related to: 10