U.S. FEDResearch
SSRN: Neacsu warns generative AI's 'the law is' answers hide the work of legal research
In an October 13, 2026 SSRN preprint, Dana Neacsu argues that when generative AI answers a legal question with "the law is," it delivers an ideologically prepackaged conclusion that hides the data pool, source selection, reading and interpretation behind it. She traces four shifts in how legal knowledge is produced and urges preserving visibility into those steps so users can reclaim epistemic agency.
Key points
- Neacsu argues AI legal answers obscure the research process: the data pool, source selection, reading and interpretation.SSRN
- The paper describes a shift from information infrastructure to epistemic infrastructure, and from source to reconstruction to authority.SSRN
- It also traces moves from informational capitalism to epistemic capitalism, and from AI regulation to the governance of meaning.SSRN
- Neacsu draws on Marx's concept of Gallerte and on intelligence fetishism to explain how AI outputs gain authority.SSRN
- She proposes keeping data pools, source selection, classification and interpretation visible so users can question rather than obey outputs.SSRN