The evidence on the effects of generative AI (GenAI) on critical thinking is mixed, with studies suggesting both potential harms and benefits depending on its implementation. Some argue that AI-driven provocations, such as questions asking for human clarification and justification, are beneficial for eliciting critical thinking. Drawing on our experience designing and evaluating two GenAI-powered tools for knowledge work, ArtBot in the domain of fine art interpretation and Privy in the domain of AI privacy, we reflect on how design decisions shape the form and effectiveness of such provocations. Our observations and user feedback suggest that domain-specific provocations, implemented through productive friction and interactions that depend on user contribution, can meaningfully support critical thinking. We present participant experiences with both prototypes and discuss how supporting critical thinking may require moving beyond static provocations toward approaches that adapt to user preferences and levels of expertise.
Figures & tables
Case Study
Domain
Theoretical Grounding
Model Type
ArtBot
Art Interpretation
Dialogic Education ( Skidmore and Murakami, 2016 )
Local, open-weight models ( Meta, 2024 )
Privy
AI Privacy Reports
Framework Grounded Provocations ( Das et al., 2022 )
Cloud-host, proprietary models ( Wei et al., 2022 )
Table 1. This table provides an overview on the theoretical grounding underlying both systems and which specific domains each one operates under.
Appendix figures & tables2 assets
Supplementary material from the paper’s appendix.
Appendix
Figure 1. ArtBot is an LLM-powered tool that challenges participants with questions to encourage them to share their own interpretations of the artwork in the shared workspace.
Figure 2. Privy is an LLM-powered tool that guides practitioners through structured privacy impact assessments to: (i) identify relevant risks in novel AI product concepts, and (ii) propose appropriate mitigations.
This intervention study investigates the use of counterarguments in writing for critical thinking by students in the context of Generative AI (GenAI). This is especially as risks of cheating and cognitive offloading exist with the use of GenAI. We presented 36 students in a particular university course with 4 carefully selected thesis statements (from a set of popular debates) to write about anyone of them. We used six established rubrics (focus, logic, content, style, correctness and reference) to conduct three human assessments (two student peer-reviews and one experienced teacher) per writeup on a 5-point Likert scale for all the qualified samples (n) of 35 submissions (after disqualifying one for irregularity). Using the same rubrics and guidelines, we also assessed the submissions using six frontier LLMs as judges. Our mixed-method design included qualitative open-ended feedback per assessment and quantitative methods. The results reveal that (1) the students' self-written counterarguments to AI-generated content contains logic, among other things, which is a key component of critical thinking, and (2) GenAI can be successfully used at scale to assess students' written work, based on clear rubrics, and these assessments generally align with human assessments as shown with Gwets AC2 inter-rater reliability values of 0.33 for all the models except one.
Tosin Adewumi, Marcus Liwicki, Foteini Simistira Liwicki +3
Machine Learning Group, EISLAB, Luleå University of Technology, Sweden.
Generative AI (GenAI) search tools are increasingly used for information seeking, yet their design tends to encourage cognitive offloading, which may lead to passive engagement, selective attention, and informational homogenization. Effective use requires metacognitive engagement to craft good prompts, verify AI outputs, and critically engage with information. We developed MetaCues, a novel GenAI-based interactive tool for information seeking that delivers metacognitive cues alongside AI responses and a note-taking interface to guide users' search and associated learning. Through an online study (N = 146), we compared MetaCues to a baseline tool without cues, across two broad search topics that required participants to explore diverse perspectives in order to make informed judgments. Preliminary findings regarding participants' search behavior show that MetaCues leads to increased confidence in attitudinal judgments about the search topic as well as broader inquiry, with the latter effect emerging primarily for the topic that was less controversial and with which participants had relatively less familiarity. Accordingly, we outline directions for future qualitative exploration of search interactions and inquiry patterns.
Anjali Singh, Karan Taneja, Zhitong Guan +1
The University of Texas at Austin, USA · Georgia Institute of Technology, USA
Generative AI makes answers easy and understanding hard, and uncritical use invites cognitive offloading. Schools still measure unaided performance, yet the real task is to produce good work with AI: framing an ill-defined task, judging the output, and steering the model toward a better result. This ability is rarely assessed in its own right; where measured, it collapses into one "prompting" score that cannot diagnose why AI use succeeds or fails. We propose CoRe-3 (Co-Reasoning), a competency model factoring productive AI use into three assessable skills we abbreviate FJS: Framing (specifying an ill-defined task before invoking AI), Judging (evaluating output for errors and unstated assumptions), and Steering (iteratively redirecting the model). Its distinguishing claim is the separation of pre-generation Framing from post-generation Steering, with Judging as the gate between. We ground the skills in theory, state five testable propositions, and instantiate them in CoReasoningLab, an open platform that presents flawed AI output and scores them independently. Over simulated learners (generated and graded by different models), the skills dissociate: each tracks its own manipulated competence while staying flat in the others, and grades become correlated when one competence is shared across all three (convergent and discriminant validity), across grader backends from two providers. Human-rater agreement and outcomes are next; we release the instrument, data, and protocol.
Alexander Apartsin, Yehudit Aperstein
Holon Institute of Technology · Afeka College of Engineering