Appendix B: Variables

Risk Vector Variables (4 Risk Vectors and Null) and Mechanisms (Binary)

In addition, binary variables of Mechanism Presence and Religion-in-AI-Role were adopted to capture the casual mechanism (how) as opposed to end-state descriptions (what) and if AI was ruled against any religious interpretative authority.

Please note that while other risk vectors used a simple binary check, Gradual Disempowerment (GD) went through an extra manual review after the LLM coding pass failed the human auditing. To capture the nuances, GD was expanded to four levels. To handle these gray areas in the final numbers, the results will be split into two thresholds:

Domains (5 Domains with 13 Sub-domains)

Three domain variables were adapted from the main societal systems (D1 Economy, D2 Culture and D3 Nation-State/Politics) as well as two domain variables relevant to theological and philosophical documents, namely D4 Humanhood and D5 AI as Conceptual Subject, relevant to theological, philosophical and normative institutional documents. These were subsequently expanded into granular sub-codes (e.g., Within D1, D1a was about macro-allocation of capital while D1b was about the metrics of human labor) by Researcher 2 deductively following a pilot analysis of approximately 100 data extracts.

Moves (6 Moves with 11 Sub-moves)

Similarly, rhetorical move variables (Moves) were deductively structured by Leech to capture the rhetorical trajectory of institutional claims. Just like the Domains, these were also subsequently expanded into granular sub-codes by Researcher 2 (e.g. Within M1 Diagnostics, M1a was neutral, M1b was positive and M1c was negative) following the pilot.

Policy Demands (5 Categories, 22 Sub-categories, 1 Other and Null)

Target Stakeholders (5)

Target stakeholders were coded by Researcher 2 to examine if the demand for actions and red lines had a specific target (i.e., tech companies, legislators, governments, regulatory organizations) instead of just the religious community or general humanity.