
PhD in Psychology | SPAR Fellow | Award-Nominated Science Communicator with 2.5M Views | EA Bristol Vice-President
ALEXANDRA
CHEUNG
I am a Psychology PhD student at the University of Bristol. Before my PhD, I founded the largest online psychoeducation social media account in HK/TW with 100K followers and 2.5 million views, translating academic theories into accessible content for the general public.
Currently, I am taking a sabbatical to focus on interdisciplinary research at the intersection of human and digital minds, including AI & Religion, Personhood of AI, AI Consciousness, and AI Psychosis.


Current & Future Projects
This includes research projects I have recently been involved in or am about to start in 2026–2027.


Is Gradual Disempowerment on the Radar of the Abrahamic Religions?


Personhood & Linguistic Emergence in Primates with AI as an Outgroup
A mixed-methods analysis of 15 authoritative religious documents (broken down into 1790 data extracts) of five major religions with hybrid human-in-the-loop coding pipeline to examine if they engage with the Gradual Disempowerment (GD) threat with early results revealing that GD did not shows up strongly in these statements.
Supervised by Dr. Gabriela Daly (Visiting Scholar of University of St Andrews and Founder of Primate Anthrozoology Research Program)
A systematic literature review how personhood is conceptualized, attributed, contested and operationalized across scientific, philosophical, anthropological, ethical, and legal frameworks as well as evaluating the criteria used to grant or deny moral and legal status to nonhuman primates. with AI used as a comparative outgroup.
Mapping the Emerging Phenomenon of AI Psychosis: A Scoping Review
A scoping review mapping how AI psychosis is defined, operationalized, and observed across peer-reviewed journals, preprints, and grey literature to identify what we actually know about this emerging phenomenon serving as the literature baseline to triangulate with experimental AI psychosis benchmarks.
Beyond Metadata: A Manual Atomic-Unit Audit of LLM Citation Faithfulness
This pilot introduces a granular auditing methodology to evaluate whether LLMs accurately reflect the source material they cite. By breaking model outputs into isolated atomic units of claims, the framework catches subtle hallucinations and misattributions that standard metadata checks miss.
How and why people change their views on phenomenal consciousness
Public Essays on AI Consciousness
A qualitative analysis of belief revision in experimental philosophy, examining how and why individuals shift their foundational views on phenomenal consciousness and digital minds. Based on in-depth interview transcripts, the study maps the psychological and argumentative triggers behind conceptual changes in consciousness theory.
Collaborating with Dr. David Leech (Honorary Research Associate in the Department of Religion and Theology at the University of Bristol)
Supervised by Dr. Maria Avramidou (Independent Researcher in AI Consciousness & Scientific Grants Advisor of Foresight Institute)
A series of interdisciplinary public essays on AI consciousness, drawing from scientific and philosophical literature to inform AI safety, governance, and policy researchers.
Research Collaborator with Chris Percy (cf-debate.com) & Asher Soryl (qualitative re-coding and thematic audit)
Collaborating with Lenz Dagohoy & Roland Pihlakas (AI Safety Researchers at Three Laws)
Fellowship and Training
Provide a short summary of your recent projects, highlighting the most important things.


AI Safety Camp Research Incubator


Technical AI Safety Course
A three-week program focusing on research scoping and creating project proposals on a topic related to AI safety.
A 10-week research internship to work on projects aimed at examining animal welfare and artificial minds.


Supervised Program for Alignment Research (SPAR)
A three-month part-time, research program working to mitigate the risks posed by advanced AI.


Primate Anthrozoology Research Program (PARP) x Sentient Futures
A six-week program focusing on scoping research topics.


AGI Strategy
A three-week program focusing on research scoping and creating project proposals on a topic related to AI safety.