Skip to main content
Quester uses AI to reduce the execution costs of research: extracting metadata, finding material, comparing a bounded corpus, surfacing patterns, and preserving context. It is not an authority on what your sources establish or what your dissertation should argue.

​
Good uses

  • Retrieve passages or sources relevant to a question
  • Summarize a source you will verify
  • Compare definitions, methods, or reported findings
  • Identify material that may contradict a proposition
  • Organize an evidence check or reading plan
  • Recover context from work you already performed

​
Work that remains yours

  • Deciding what a source means in context
  • Judging whether evidence warrants a claim
  • Choosing theoretical and methodological commitments
  • Resolving ambiguity or disagreement
  • Forming the dissertation’s argument
  • Writing prose that represents your scholarship

​
A verification discipline

For consequential AI-assisted output:
  1. Constrain the literature and research-question scope.
  2. Ask for citations or source references.
  3. Open the original source.
  4. Check that the cited passage exists and supports the wording.
  5. Record your own interpretation separately.
  6. Cite the source, never the generated answer as a substitute for evidence.

​
Protect uncertainty

The Advisor’s Source basis notice distinguishes retrieved excerpts from saved records or analysis. Retrieved excerpts cover only part of the available material; a records-based answer retrieved no new source passages. Selecting sources focuses the request but does not establish that every selected document was read. An answer marked unfinished may contain partial text or blocks. Use Advisor recovery before treating it as a completed analysis. Generated answers tend to sound more settled than the underlying literature. Preserve uncertainty, boundary conditions, missing evidence, and genuine disagreement. If the sources do not support an answer, “not established by this corpus” is a successful result.

​
Follow institutional rules

Your university, funder, journal, or research ethics process may impose additional rules for AI use, confidential material, authorship, disclosure, and data handling. Those rules take precedence over convenience.
The standard is not whether an AI response sounds plausible. The standard is whether you can defend the resulting claim from the underlying evidence.
Last modified on September 6, 2026