Sabino MarquezWriting
Back to manuscripts

Sabino Marquez · Research methodology · AI authority

Bounded Research Infrastructure

Who decided what your research could claim?

A research protocol for keeping AI authority visible and subject to human judgment

Manuscript Progress

Stage 6 of 7
  1. Ideation — Complete
  2. Outline — Complete
  3. First draft — Complete
  4. Revision draft — Complete
  5. Copy edit — Complete
  6. QA /
    Pre-release
    — Current stage
  7. Publish — Not yet reached

An AI summary can remove a mechanism that the source depends on. That summary can then shape a literature review or settle the structure of an argument. The effect may survive after every generated sentence has been discarded. A researcher still needs to account for how the final claim took its form.

Bounded Research Infrastructure examines the authority AI acquires through evidence selection, interpretation, structure, and acceptance. Sabino Marquez proposes a bounded protocol stack that keeps those decisions inspectable throughout the research workflow. Researchers retain responsibility for the model, the evidence, and admission of material into the research corpus.

BRI develops and follows encounters in which useful AI outputs required rejection, transformation, or a return to the source. Provenance includes those decisions because they explain why the accepted claim survived.

Researchers, reviewers, and editors gain an account they can interrogate: what the AI was allowed to determine, what the researcher changed, and why the resulting claim was accepted. The paper also identifies requirements for tools that can preserve that account. Its proposed authority controls provide a basis for further empirical evaluation.

What you’ll discover
  • Locate AI authority in a workflow even when generated prose does not survive
  • Examine the human decisions that connect source material to an accepted claim
  • Understand why rejection and transformation belong in research provenance
  • Identify what research tools need to preserve for meaningful review
Sabino Marquez smiling in a blue sweater against a brick background.

About the author

Sabino Marquez

Sabino Marquez is a researcher and the co-founder of Trustable. Bounded Research Infrastructure develops an account of AI authority through his own source-handling and model-formalization practice.

Explore the practice

Reading And Research

For questions about reading the paper or examining the protocol in your research practice, contact Sabino.

Available only through a Trustable engagement.

Ask about this paper