How this research uses Claude
A plain-English version of the FT-series LLM disclosure policy
This page explains how the Friction Theory research programme uses Claude (a large language model by Anthropic), and how that is acknowledged in each paper. The same short paragraph appears in every paper in the series.
Use of Claude
I develop this work in dialog with Claude, a large language model by Anthropic (multiple model versions over the course of the programme). I use it as a discussion, literature-search, and drafting partner. The theoretical claims, the empirical design, and the responsibility are mine.
Authorship
I am the sole author of every paper in the FT series. Per current ICMJE, COPE, and WAME guidelines, large language models cannot be authors and are therefore not listed as authors or co-authors. I take full responsibility for everything published under my name.
Acknowledgement in each paper
Every paper in the FT series includes the following standard paragraph in its Acknowledgements section, placed late in the paper, typically just before the References:
Acknowledgements
The research, analysis, and writing of this paper were conducted by the author in collaboration with Claude Code (Anthropic) as a discussion, literature-search, and drafting partner. The theoretical claims, the empirical design, and responsibility for any errors are the author's alone.
The paragraph is identical across all papers. There is no tier-system or per-paper variation. Different papers involve Claude to different degrees in practice, but the in-paper text is uniform.
Why placed late
My own behavioural-friction research (Papers 0, 1) indicates that prominently-placed evaluative signals near the start of a paper function as pre-emptive friction-signals that bias subsequent reading. Late placement lets the argument stand on its own merits when readers first encounter it, with the acknowledgement following. The paragraph is short and clear; it is not buried, and it is not decorative.
Supplementary materials
Supplemental dialog logs are available on reasonable academic request. Redactions are limited to unrelated material, commercial proprietary information about Claude's underlying model, and content not relevant to the research.