Student work
Publications
Every accepted manuscript is hosted here permanently. Papers and mini-papers are listed by date within their section; outstanding work is occasionally noted as an editor's selection but is not ranked.
No publications match this filter yet.
Papers
What counts as a paper →-
Anthropomorphism of AI and Use Intensity as Mediating Variables in Adolescents’ Expectations for Human Relationships
Adolescence is a developmental period in which internal working models of relationships are actively revised, making teens disproportionately sensitive to the relational environments they inhabit. AI companions, which are much more available, consistently validating, and conflict-avoidant than humans, now make up a significant portion of adolescents’ social life. This paper asks and contributes to answering the question of whether AI use predicts shifts in adolescents’ expectations from their other relationships across four dimensions—availability, emotional validation, conflict tolerance, and honesty. Furthermore, the paper seeks to determine whether this observation is influenced by the anthropomorphization of AIs. By investigating attachment theory, research on AI-mediated social cognition, literature on chatbots and human relationships, and a Canadian high school’s student population via a survey, this cross-sectional mixed-methods study (1) formulates five hypotheses (H1-H5) and (2) creates its own measure of AI anthropomorphism by extending the RoPE scale. It then (3) tests hypotheses among a student body (N = 23) and (4) builds plausible explanations behind important findings. Results show a striking split between users and non-users: users rated their expectations for AI’s availability indistinguishably from family members, while non-users rated it the lowest out of any counterpart. Additionally, users' expectations of AI were considerably higher than non-users. Despite significant changes in expectations for AI, however, there was only a negligible change with expectations for human counterparts between users and non-users. These findings may help inform digital literacy curricula, parental guidance around AI use, and the ethical design of conversational AI systems accessible to adolescents.
Reviewed and deemed ready for publication by Dr. Jonthon Coulson. Methodology approved by OKM faculty prior to the founding of OSRI. The author is OSRI's Founding Editor and took no part in the decision on this paper.
Mini-papers
What counts as a mini-paper →No mini-papers yet. A mini-paper is one to three paragraphs: a single claim, the evidence for it, and one limitation. It is the quickest way to have real work published and reviewed.
Student-produced datasets
Not yet open. OSRI plans to archive cleaned, de-identified datasets produced by student authors so that future students can build on them. Until the archive opens, authors may attach a data appendix to a paper submission; see the guidelines for what may be shared.