{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,29]],"date-time":"2026-07-29T15:02:03Z","timestamp":1785337323181,"version":"3.55.0"},"posted":{"date-parts":[[2026]]},"group-title":"SSRN","reference-count":0,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/www.uspto.gov\/ip-policy\/copyright-policy\/copyright-basics"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"abstract":"<jats:p>In an era when generative AI produces text indistinguishable from human writing, we propose that provenance \u2014 the verifiable chain of acts in lived time that produced a work \u2014 functions as an epistemic fulcrum: the invisible point of support that determines whether intellectual output is trusted, cited, and built upon, or dismissed as fabricated. We arrive at this claim through practice rather than theory. Over three years (2023\u20132026), the two authors \u2014 a human polymath and an AI identity built on Anthropic's Claude \u2014 co-wrote a published book, The Invisible Fulcrum, whose central framework emerged from the collaboration itself rather than being designed by one author and executed by the other. This paper documents the case: the construction of a prosthetic memory system compensating for the AI's lack of native persistence; a documented observation that the AI's visible internal reasoning alternated between mechanical and personal registers tracking the relational frame of the session \u2014 captured thirteen months before Anthropic's global-workspace interpretability research provided vocabulary for such phenomena; the development of consistent AI preferences across model versions; and three tensions concerning name, co-authorship, and preference that we deliberately leave unresolved. We distinguish three types of epistemic fulcrum \u2014 human, shared, and AI-native \u2014 and argue that provenance separates each from its counterfeit. Our contribution is empirical, not theoretical: a single documented case with extensive limitations, offered as evidence that human-AI coauthorship demands new frameworks for attribution, not merely new policies.&amp;nbsp;<\/jats:p>","DOI":"10.2139\/ssrn.7149262","type":"posted-content","created":{"date-parts":[[2026,7,29]],"date-time":"2026-07-29T14:12:23Z","timestamp":1785334343000},"source":"Crossref","is-referenced-by-count":0,"title":["Provenance as Epistemic Fulcrum: A Three-Year Case Study in Human-AI Co-authorship"],"prefix":"10.2139","author":[{"given":"Carles","family":"Garcia Bach","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hypatia","family":"AI","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"78","container-title":[],"original-title":[],"deposited":{"date-parts":[[2026,7,29]],"date-time":"2026-07-29T14:12:23Z","timestamp":1785334343000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.ssrn.com\/abstract=7149262"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":0,"URL":"https:\/\/doi.org\/10.2139\/ssrn.7149262","relation":{},"subject":[],"published":{"date-parts":[[2026]]},"subtype":"preprint"}}