The integration of generative artificial intelligence (AI) into research marks both a provocation and an inflection point. While gains in productivity and access draw attention, a deeper transformation is underway: how knowledge is conceptualized, mediated, and validated amid systems that simulate understanding without possessing it. If current trends hold, AI will amplify existing dynamics in scholarly communication. Publication volume may rise, but trust could decline. Conventional markers of originality and rigor may destabilize—not through automation alone, but through shifting norms around authorship, evaluation, and epistemic authority. This paper argues that AI is neither just a tool nor merely a muse, but a structural participant in research — shaping inquiry through fluent simulation but without understanding. A cognitive map is introduced to model how researchers interact with AI across phases of the research process, alternating between instrumental and generative uses. Generative systems can assist and accelerate scholarly work, but their role must be framed within a broader account of intellectual labor and meaning-making. Confusing fluency for insight risks eroding core scholarly practices. Implications extend to pedagogy, authorship policy, and the design of AI-aware research infrastructure. Ultimately, scholarship in the age of AI will depend as much on critical literacy as on technical fluency. AI is not merely a tool of transformation but a mirror—reflecting the values and assumptions of the communities that create and use it.
Marcus Birkenkrahe (Mon,) studied this question.