Table 4 Points of friction with AI... | Emerald Publishing

Table 4

Points of friction with AI writing tools

Documented rhetorical feature	Manifestation of AI conflict	Illustrative participant quote
Elaboration as Eloquence	AI tools systematically flagging elaborated prose as “wordy,” “hard to read,” or suggesting cuts that dismantle the core rhetorical strategy	“Grammarly constantly tells me to reduce my sentences. It wants me to cut the very parts that, to me, explain the nuance. It feels like arguing with a machine that doesn't understand depth.” (Participant 21)
Context-Before-Claim	The AI's “clarity” and “conciseness” algorithms cannot parse the logical function of context-building paragraphs, recommending their deletion or severe shortening	“I wrote a beautiful introduction setting up the philosophical problem. ChatGPT summarized it into one blunt sentence and deleted the rest. It got the ‘point’ but killed the argument's soul.” (Participant 9)
Inductive Reasoning	AI tools, trained on IMRaD structures, fail to recognize the coherence of a chronological, discovery-based narrative, suggesting reorganizations that impose a rigid, deductive frame	“When I paste my results section, the AI always suggests I start with the final outcome. It doesn't understand that the process of finding it is the story. It just reorganizes my logic into its own template.” (Participant 30)

Documented rhetorical feature	Manifestation of AI conflict	Illustrative participant quote
Elaboration as Eloquence	AI tools systematically flagging elaborated prose as “wordy,” “hard to read,” or suggesting cuts that dismantle the core rhetorical strategy	“Grammarly constantly tells me to reduce my sentences. It wants me to cut the very parts that, to me, explain the nuance. It feels like arguing with a machine that doesn't understand depth.” (Participant 21)
Context-Before-Claim	The AI's “clarity” and “conciseness” algorithms cannot parse the logical function of context-building paragraphs, recommending their deletion or severe shortening	“I wrote a beautiful introduction setting up the philosophical problem. ChatGPT summarized it into one blunt sentence and deleted the rest. It got the ‘point’ but killed the argument's soul.” (Participant 9)
Inductive Reasoning	AI tools, trained on IMRaD structures, fail to recognize the coherence of a chronological, discovery-based narrative, suggesting reorganizations that impose a rigid, deductive frame	“When I paste my results section, the AI always suggests I start with the final outcome. It doesn't understand that the process of finding it is the story. It just reorganizes my logic into its own template.” (Participant 30)