Key answer
Roles can change tone, analytical perspective and accuracy; these effects must be distinguished. Sounding like an expert is easy to notice in wording; a correct judgment requires separate verification.
Full explanation
*Principled Personas* compared role prompts across nine open-weight models and 27 knowledge and reasoning tasks. Some combinations genuinely improved accuracy, others made no significant difference, and some declined. Roles can therefore do more than embellish language, but a role's name does not predict greater accuracy in a particular case. See Q001 for the fuller comparison. Paper
Split the check into three parts. For an editor's perspective, check whether it identifies specific structure, wording and reader-comprehension problems. For plain language, inspect the expression. For factual accuracy, compare with original sources, data or reference answers. Technical vocabulary cannot substitute for either clear expression or factual verification.
For example:
Review this passage as a textbook editor. Separately identify misused terms, missing reasoning steps and places beginners may misunderstand. Provide verifiable support for factual judgments; label suggestions based only on stylistic preference separately.
This expands the professional title into observable work. It helps acceptance checks, but the AI's supporting evidence still needs inspection.