Attachment to LLMs exhibits cruel optimism: desired efficiency collides with the vigilance and expertise students do not yet possess¶

Claim · All claims
Evidence · 1 study · associational r1 · q1
Currency · newest study 2025 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q1 i? The chapter argues, as a theoretical interpretation, that students' attachment to LLM promises is a cruel optimism, because the perceived benefits depend on the very skills, vigilance and expertise students are still developing. → Kudina 2025

Evidence¶

Kudina 2025¶

Kudina, O. (2025). Using AI in engineering education: a balancing act, driven by clear purpose. Draft version, to appear in The Routledge Handbook of the Philosophy of Engineering, 2nd ed. https://arxiv.org/abs/2606.16626 doi:10.48550/arxiv.2606.16626

q1 · i? · associational · r1

Theoretical argument drawing on Berlant (2020), grounded in the questionnaire findings on verification burden and conditional positive evaluation; this is the author's interpretation, not a separately tested result.

"the perceived or desired efficiency collides with the actual required vigilance and output evaluation. Here, the promises of ease of use and time-saving do not match the additional, and at times significant, cognitive, emotional, and moral labor"

Discussion¶

Related Claims¶

• Fine-tuned PLMs measure high-inference teaching practices better when the variable requires less pedagogical expertise, matching human-rater agreement on lexical variables but degrading on inference-heavy ones — related
• Fully inductive LLM codebook development risks importing unexamined sensitizing concepts, such as folk theories and scientific misconceptions, from the model's training data — related
• Five common AI misconceptions debunked: AI does not think like humans, create original ideas, replace teachers, or guarantee accuracy, and well-designed AI use need not lower rigor — related
• The predominant student evaluation of LLMs (42%) is positive but conditional on the student's own verification work — related
• Design factors moderating AI's effect on learning matter only through whether they engage the learner's evaluation — related
• LLM output can be misleading or erroneous because a stochastic element in text generation sometimes selects less likely continuations, so it requires continuous supervision and critical evaluation by domain experts. — related
• Optimizing prompts can vastly improve the accuracy of LLM outputs, but critical evaluation of the resulting outputs still requires domain expertise. — related
• Calibrated epistemic vigilance, given adequate prior knowledge, is the binding constraint on productive augmentation with AI — a broader claim this one bears on