pith. sign in

arxiv: 2606.03095 · v1 · pith:EVTXWYPXnew · submitted 2026-06-02 · 💻 cs.HC

AI Assistance for Discretionary Work: Increasing Feedback Provision in Higher Education

Romina Mahinpei , Victoria Dean , Ruth Fong , Lydia T. Liu , Manoel Horta Ribeiro This is my paper

Pith reviewed 2026-06-28 08:52 UTC · model grok-4.3

classification 💻 cs.HC
keywords AI assistancediscretionary workfeedback provisionhigher educationrandomized field experimentteaching assistants
0
0 comments X

The pith

AI-generated feedback drafts increase the share of student submissions receiving comments by 10.8 percentage points.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper tests whether AI can raise participation in discretionary but valuable tasks that people often intend to do yet skip, using personalized feedback on student work as the setting. In a randomized field experiment inside a machine-learning course, teaching assistants who received AI drafts after grading provided feedback on more submissions and wrote longer comments than those without drafts. Student ratings of usefulness stayed the same and time per character did not fall, indicating that the extra output did not come at the expense of quality or effort. The authors interpret the drafts as editable starting points that lower the barrier to beginning the task while leaving final control with the human TA.

Core claim

In the randomized experiment, student submissions assigned to the AI-draft condition received feedback 10.8 percentage points more often and the feedback was 39.8 characters longer on average, with no reduction in student-perceived usefulness or time spent per character written; the authors conclude that the drafts served as scaffolds that reduced initiation costs rather than replacing human work.

What carries the argument

AI-generated draft feedback provided to TAs after they grade, which TAs may adopt, edit, or discard at their discretion.

If this is right

  • AI assistance can raise completion rates for discretionary tasks without reducing the quality of the final output.
  • Human users retain control because they decide whether and how to use the drafts.
  • The increase occurs mainly through more frequent initiation of the task rather than through faster completion of each instance.
  • The same pattern may apply to other optional but beneficial professional activities where people intend to act but often do not.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Similar draft-based assistance could be tested in settings outside education such as code review or performance evaluations where feedback is discretionary.
  • Longer-term studies could check whether repeated exposure to drafts changes TAs' own writing habits or reduces the need for assistance over time.
  • The design leaves open whether the benefit depends on the quality or style of the specific AI model used to generate the drafts.

Load-bearing premise

The measured rise in feedback occurs because the drafts lower the cost of starting the writing task rather than because of unmeasured differences between the TAs in the two groups or because of features unique to the machine-learning course.

What would settle it

A replication in a different course or with a different AI model that produces no statistically significant increase in feedback rates or lengths would indicate that the effect is not generalizable beyond the tested setting.

Figures

Figures reproduced from arXiv: 2606.03095 by Lydia T. Liu, Manoel Horta Ribeiro, Romina Mahinpei, Ruth Fong, Victoria Dean.

Figure 1
Figure 1. Figure 1: Study overview and main findings. After grading each question submission, teaching assistants (TAs) are randomly assigned to receive either an AI-assisted feedback draft (treatment) or no feedback draft (control). Feedback drafts can be used, edited, or ignored completely at the TA’s discretion. We find that AI-assisted feedback increased feedback provision (+10.8 percentage points) and length (+39.8 chara… view at source ↗
Figure 2
Figure 2. Figure 2: Overview of the formative study. Two parallel components informed the design of our AI-assisted feedback intervention: (1) instructor-informed intervention design, which included iterative prompt engineer￾ing and workflow design with course instructors, and (2) a TA evaluation for model selection, in which TAs ranked feedback drafts generated by four candidate LLMs and an option to write feedback from scra… view at source ↗
Figure 3
Figure 3. Figure 3: Integration of AI-assisted feedback drafts into the grading workflow. We develop a lightweight, LLM-backed Chrome extension that surfaces feedback drafted after grading via a “Done Grading” button (A). In the control condition, TAs proceed without assistance (B), whereas in the treatment condition, they are shown a personalized AI-assisted feedback draft (C). Both conditions included a post-grading message… view at source ↗
Figure 4
Figure 4. Figure 4: Question-level randomization with rotating assignment. Students were randomly assigned to Group A or B at the start of each assignment. Treatment and control conditions alternated across questions within an assignment: for Question 1, Group A received treatment and Group B control; for Question 2, the assignment reversed; and this rotation continued for subsequent questions. Each question was graded by a s… view at source ↗
Figure 5
Figure 5. Figure 5: Temporal dynamics of AI-assisted feedback provision and feedback draft usage over the semester. (A) AI assistance consistently increases feedback provision across all four homework assignments, though feedback provision declines over time in both conditions as the semester progresses. (B) Template usage is initially high but decreases over later assignments, suggesting that AI-assisted drafts may primarily… view at source ↗
read the original abstract

AI systems increasingly shape human workflows by generating intermediate artifacts that users can adopt, revise, or ignore. While prior work has shown that AI assistance can improve the efficiency and accuracy of required tasks, less is known about whether it can increase participation in discretionary but beneficial work that users often intend to perform but frequently skip. We study this question in the context of personalized feedback provision in higher education, a pedagogically valuable but often optional practice. We conduct a mixed-methods study combining a randomized field experiment and qualitative interviews in a 300-level machine learning course with n=11 teaching assistants (TAs) and n=88 students. Student submissions were randomly assigned to either (1) a treatment condition where TAs received AI-assisted feedback drafts after grading or (2) a control condition without drafts. TAs remained fully in control and could use, edit, or ignore drafts at their discretion. We find that AI-assisted feedback significantly increases feedback provision (+10.8 percentage points, SE=1.1, p<0.001) and feedback length (+39.8 chars, SE=3.45, p<0.001) without negatively affecting student usefulness ratings or reducing time per character. Qualitative findings suggest that AI-assisted drafts function as editable scaffolds that lower barriers to initiating feedback rather than reducing overall effort. Our findings highlight AI's promise for discretionary but beneficial tasks: increasing work that might otherwise go undone while preserving human control over final outcomes.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 2 minor

Summary. The paper reports results from a mixed-methods study (randomized field experiment plus qualitative interviews) in a 300-level machine learning course (n=11 TAs, n=88 students). Student submissions were randomly assigned to a treatment condition in which TAs received AI-generated feedback drafts after grading or to a control condition without drafts; TAs retained full discretion to use, edit, or ignore the drafts. The central empirical claims are that the AI drafts increase the probability of feedback provision by 10.8 percentage points (SE=1.1, p<0.001) and increase feedback length by 39.8 characters (SE=3.45, p<0.001), with no detectable negative effects on student-rated usefulness or time per character. Qualitative interviews interpret the drafts as editable scaffolds that lower initiation costs rather than substitute for effort.

Significance. If the reported effects hold, the work supplies field-experimental evidence that AI assistance can raise completion rates for discretionary but pedagogically valuable tasks while preserving human control over final outputs. The submission-level randomization, pre-registered-style reporting of effect sizes with SEs and p-values, and alignment between quantitative and qualitative results are notable strengths that support causal attribution to the availability of drafts. The findings are relevant to workflow design in education and other domains where beneficial optional work is frequently omitted.

minor comments (2)
  1. [Methods] The description of the randomization procedure (submission-level assignment) is clear, but a brief statement confirming that the randomization was performed after grading but before draft generation would eliminate any residual ambiguity about timing.
  2. [Qualitative Findings] The qualitative section would benefit from one additional sentence explicitly linking the interview themes back to the quantitative outcome measures (e.g., initiation barrier vs. effort reduction).

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive assessment of the manuscript, accurate summary of the study design and findings, and recommendation to accept. We appreciate the recognition of the submission-level randomization, pre-registered-style reporting, and alignment between quantitative and qualitative evidence.

Circularity Check

0 steps flagged

No circularity: empirical RCT results with no derivation chain

full rationale

The paper reports causal effects from a randomized field experiment (student submissions assigned to AI-draft vs. control) and qualitative interviews. Central claims are direct measurements (+10.8 pp feedback provision, +39.8 chars length) against an external control group, with randomization balancing confounders. No equations, fitted parameters renamed as predictions, self-definitional constructs, or load-bearing self-citations appear in the provided text. The derivation chain is absent; results are externally falsifiable via the experiment design itself.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The claim rests on standard assumptions of randomized experiments (exchangeability under random assignment) and on the validity of the feedback-provision metric; no free parameters, invented entities, or ad-hoc axioms are introduced.

axioms (2)
  • standard math Random assignment produces comparable treatment and control groups on average.
    Invoked to interpret the 10.8 pp difference as causal.
  • domain assumption Feedback provision and length are measured consistently across conditions.
    Required for the reported effect sizes to be comparable.

pith-pipeline@v0.9.1-grok · 5803 in / 1307 out tokens · 26222 ms · 2026-06-28T08:52:21.563544+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

64 extracted references · 41 canonical work pages

  1. [1]

    Faria Ahmed, Nabila Bouali, and Marion Gerhold. 2024. Teaching Assistants as Assessors: An Experience-Based Narrative. InProceedings of the 16th International Conference on Computer Supported Education (CSEDU). SCITEPRESS – Science and Technology Publications, 315–323

    2024

  2. [2]

    2014.The Invisible Work of Nurses: Hospitals, Organisation and Healthcare(1 ed.)

    Davina Allen. 2014.The Invisible Work of Nurses: Hospitals, Organisation and Healthcare(1 ed.). Routledge. https: //doi.org/10.4324/9781315857794

    work page doi:10.4324/9781315857794 2014

  3. [3]

    Saleema Amershi, Maya Cakmak, William Bradley Knox, and Todd Kulesza. 2014. Power to the people: The role of humans in interactive machine learning.AI magazine35, 4 (2014), 105–120

    2014

  4. [4]

    Bennett, Kori Inkpen, Jaime Teevan, Ruth Kikin-Gil, and Eric Horvitz

    Saleema Amershi, Dan Weld, Mihaela Vorvoreanu, Adam Fourney, Besmira Nushi, Penny Collisson, Jina Suh, Shamsi Iqbal, Paul N. Bennett, Kori Inkpen, Jaime Teevan, Ruth Kikin-Gil, and Eric Horvitz. 2019. Guidelines for Human-AI Interaction. InProceedings of the 2019 CHI Conference on Human Factors in Computing Systems(Glasgow, Scotland Uk) (CHI ’19). Associa...

    work page doi:10.1145/3290605.3300233 2019

  5. [5]

    John W. Budd. 2016.2. The Eye Sees What the Mind Knows: The Conceptual Foundations of Invisible Work. University of California Press, Berkeley, 28–46. https://doi.org/doi:10.1525/9780520961630-004

    work page doi:10.1525/9780520961630-004 2016

  6. [6]

    Julia Cambre, Scott Klemmer, and Chinmay Kulkarni. 2018. Juxtapeer: Comparative Peer Review Yields Higher Quality Feedback and Promotes Deeper Reflection. InProceedings of the 2018 CHI Conference on Human Factors in Computing Systems(Montreal QC, Canada)(CHI ’18). Association for Computing Machinery, New York, NY, USA, 1–13. https://doi.org/10.1145/317357...

    work page doi:10.1145/3173574.3173868 2018

  7. [7]

    Yulu Cui, Weimiao He, Xingke Du, Manlin Zeng, and Dongping Liu. 2025. The Impact of AI Writing Assistants on Academic Writing Performance:.International Journal of Distance Education Technologies23, 1 (2025). https: //doi.org/10.4018/IJDET.391326

    work page doi:10.4018/ijdet.391326 2025

  8. [8]

    Arlene Kaplan Daniels. 1987. Invisible Work.Social Problems34, 5 (1987), 403–415. https://doi.org/10.2307/800538

    work page doi:10.2307/800538 1987

  9. [9]

    Marjorie L. DeVault. 2014. Mapping Invisible Work: Conceptual Tools for Social Justice Projects.Sociological Forum29, 4 (2014), 775–790. http://www.jstor.org/stable/43654145

    arXiv 2014

  10. [10]

    Doe, Katherine J

    Susan R. Doe, Katherine J. Gingerich, and Tara L. Richards. 2013. An Evaluation of Grading and Instructional Feedback Skills of Graduate Teaching Assistants in Introductory Psychology.Teaching of Psychology40, 4 (2013), 274–280. https://doi.org/10.1177/0098628313501039

    work page doi:10.1177/0098628313501039 2013

  11. [11]

    Mica R Endsley. 2018. Automation and situation awareness. InAutomation and human performance. CRC Press, 163–181

    2018

  12. [12]

    Juan Escalante, Austin Pack, and Alex Barrett. 2023. AI-generated feedback on writing: insights into efficacy and ENL student preference.International Journal of Educational Technology in Higher Education20, 1 (2023), 57. https: //doi.org/10.1186/s41239-023-00425-2

    work page doi:10.1186/s41239-023-00425-2 2023

  13. [13]

    Emily Faulconer, John Griffith, and Amy Gruss. 2022. The impact of positive feedback on student outcomes and perceptions.Assessment & Evaluation in Higher Education47, 2 (2022), 259–268. https://doi.org/10.1080/02602938. 2021.1910140

    work page doi:10.1080/02602938 2022

  14. [14]

    Fleur, Max Marshall, Miguel Pieters, Natasa Brouwer, Gerrit Oomens, Angelos Konstantinidis, Koos Winnips, Sylvia Moes, Wouter van den Bos, Bert Bredeweg, and Erwin A

    Damien S. Fleur, Max Marshall, Miguel Pieters, Natasa Brouwer, Gerrit Oomens, Angelos Konstantinidis, Koos Winnips, Sylvia Moes, Wouter van den Bos, Bert Bredeweg, and Erwin A. van Vliet. 2023. IguideME: Supporting Self-Regulated Learning and Academic Achievement with Personalized Peer-Comparison Feedback in Higher Education.Journal of Learning Analytics1...

    work page doi:10.18608/jla.2023.7853 2023

  15. [15]

    Anupriya Sharma Ghai, Vandna Rawat, Vishan Kumar Gupta, and Kapil Ghai. 2024. Artificial Intelligence in System and Software Engineering for Auto Code Generation. In2024 International Conference on Electrical Electronics and Computing Technologies (ICEECT), Vol. 1. 1–5. https://doi.org/10.1109/ICEECT61758.2024.10738945

    work page doi:10.1109/iceect61758.2024.10738945 2024

  16. [16]

    Evelyn Nakano Glenn. 2000. Creating a Caring Society.Contemporary Sociology29, 1 (2000), 84–94. https://doi.org/10. 2307/2654934

    2000

  17. [17]

    Jonathan Grudin. 1988. Why CSCW applications fail: problems in the design and evaluationof organizational interfaces. InProceedings of the 1988 ACM Conference on Computer-Supported Cooperative Work(Portland, Oregon, USA)(CSCW ’88). Association for Computing Machinery, New York, NY, USA, 85–93. https://doi.org/10.1145/62266.62273

    work page doi:10.1145/62266.62273 1988

  18. [18]

    Ashish Gurung, Jionghao Lin, Jordan Gutterman, Danielle R Thomas, Alex Houk, Shivang Gupta, Emma Brunskill, Lee Branstetter, Vincent Aleven, and Kenneth Koedinger. 2025. Human Tutoring Improves the Impact of AI Tutor Use on Learning Outcomes. InArtificial Intelligence in Education: 26th International Conference, AIED 2025, Palermo, Italy, July 22–26, 2025...

    work page doi:10.1007/978-3-031-98459-4_28 2025

  19. [19]

    John Hattie and Helen Timperley. 2007. The Power of Feedback.Review of Educational Research77, 1 (2007), 81–112. https://doi.org/10.3102/003465430298487

    work page doi:10.3102/003465430298487 2007

  20. [20]

    Erin Hatton. 2017. Mechanisms of Invisibility: Rethinking the Concept of Invisible Work.Work, Employment and Society31, 2 (2017), 336–351. 22 Mahinpei et al

    2017

  21. [21]

    Michael Henderson, Tracii Ryan, and Michael Phillips. 2019. The challenges of feedback in higher education.Assessment & Evaluation in Higher Education44, 8 (2019), 1237–1252. https://doi.org/10.1080/02602938.2019.1599815

    work page doi:10.1080/02602938.2019.1599815 2019

  22. [22]

    Eric Horvitz. 1999. Principles of mixed-initiative user interfaces. InProceedings of the SIGCHI Conference on Human Factors in Computing Systems(Pittsburgh, Pennsylvania, USA)(CHI ’99). Association for Computing Machinery, New York, NY, USA, 159–166. https://doi.org/10.1145/302979.303030

    work page doi:10.1145/302979.303030 1999

  23. [23]

    Lucas Jasper Jacobsen and Kira Elena Weber. 2025. The Promises and Pitfalls of Large Language Models as Feedback Providers: A Study of Prompt Engineering and the Quality of AI-Driven Feedback.AI6, 2 (2025). https://doi.org/10. 3390/ai6020035

    2025

  24. [24]

    Ziwei Ji, Nayeon Lee, Rita Frieske, Tiezheng Yu, Dan Su, Yan Xu, Etsuko Ishii, Ye Jin Bang, Andrea Madotto, and Pascale Fung. 2023. Survey of Hallucination in Natural Language Generation.ACM Comput. Surv.55, 12, Article 248 (March 2023), 38 pages. https://doi.org/10.1145/3571730

    work page doi:10.1145/3571730 2023

  25. [25]

    Qinjin Jia, Jialin Cui, Ruijie Xi, Chengyuan Liu, Parvez Rashid, Ruochi Li, and Edward Gehringer. 2024. On Assessing the Faithfulness of LLM-Generated Feedback on Student Assignments. InProceedings of the 17th International Conference on Educational Data Mining. International Educational Data Mining Society, 491–499. https://doi.org/10.5281/zenodo. 12729868

    work page doi:10.5281/zenodo 2024

  26. [26]

    Briese, Marius Jacobs, Niclas Dern, Niels Glodny, Simon Jacobs, and Samuel Leßmann

    Annette Kinder, Fiona J. Briese, Marius Jacobs, Niclas Dern, Niels Glodny, Simon Jacobs, and Samuel Leßmann. 2025. Effects of adaptive feedback generated by a large language model: A case study in teacher education.Computers and Education: Artificial Intelligence8 (2025), 100349. https://doi.org/10.1016/j.caeai.2024.100349

    work page doi:10.1016/j.caeai.2024.100349 2025

  27. [27]

    Yong Ming Kow and Waikuen Cheng. 2018. Complimenting Invisible Work: Identifying Hidden Employee Contributions through a Voluntary, Positive, and Open Work Review System.Proc. ACM Hum.-Comput. Interact.2, CSCW, Article 96 (Nov. 2018), 22 pages. https://doi.org/10.1145/3274365

    work page doi:10.1145/3274365 2018

  28. [28]

    Kraut, Paul Resnick, Sara Kiesler, Moira Burke, Yan Chen, Niki Kittur, Joseph Konstan, Yuqing Ren, and John Riedl

    Robert E. Kraut, Paul Resnick, Sara Kiesler, Moira Burke, Yan Chen, Niki Kittur, Joseph Konstan, Yuqing Ren, and John Riedl. 2012.Building Successful Online Communities: Evidence-Based Social Design. The MIT Press. https: //doi.org/10.7551/mitpress/8472.001.0001

    work page doi:10.7551/mitpress/8472.001.0001 2012

  29. [29]

    John D Lee and Katrina A See. 2004. Trust in automation: Designing for appropriate reliance.Human factors46, 1 (2004), 50–80

    2004

  30. [30]

    Chin-Yew Lin. 2004. ROUGE: A Package for Automatic Evaluation of Summaries. InText Summarization Branches Out. Association for Computational Linguistics, Barcelona, Spain, 74–81. https://aclanthology.org/W04-1013/

    2004

  31. [31]

    Xinyi Lu, Kexin Phyllis Ju, Mitchell Dudley, Larissa Sano, and Xu Wang. 2026. AI-Mediated Feedback Improves Student Revisions: A Randomized Trial with FeedbackWriter in a Large Undergraduate Course. arXiv:2602.16820 [cs.HC] https://arxiv.org/abs/2602.16820

    arXiv 2026

  32. [32]

    Romina Mahinpei, Sofiia Druchyna, and Xinran Bi. 2026. CNPE: A Framework of Challenges & Needs in Proof Education. InProceedings of the 57th ACM Technical Symposium on Computer Science Education V.2(USA). Association for Computing Machinery, New York, NY, USA, 1435–1436. https://doi.org/10.1145/3770761.3777193

    work page doi:10.1145/3770761.3777193 2026

  33. [33]

    Romina Mahinpei, Sofiia Druchyna, and Manoel Horta Ribeiro. 2026. When LLMs Help – and Hurt – Teaching Assistants in Proof-Based Courses. arXiv:2602.23635 [cs.HC] https://arxiv.org/abs/2602.23635

    arXiv 2026

  34. [34]

    The Power of Feedback

    Luke Mandouit and John Hattie. 2023. Revisiting “The Power of Feedback” from the perspective of the learner.Learning and Instruction84 (2023), 101718. https://doi.org/10.1016/j.learninstruc.2022.101718

    work page doi:10.1016/j.learninstruc.2022.101718 2023

  35. [35]

    John Meluso, Amanda Casari, Katie McLaughlin, and Milo Z Trujillo. 2025. Invisible Labor in Open Source Software Ecosystems.Proc. ACM Hum.-Comput. Interact.9, 7, Article CSCW236 (Oct. 2025), 32 pages. https://doi.org/10.1145/ 3757417

    2025

  36. [36]

    Liebenow, Marlene Steinbach, Andrea Horbach, and Johanna Fleckenstein

    Jennifer Meyer, Thorben Jansen, Ronja Schiller, Lucas W. Liebenow, Marlene Steinbach, Andrea Horbach, and Johanna Fleckenstein. 2024. Using LLMs to bring evidence-based feedback into the classroom: AI-generated feedback increases secondary students’ text revision, motivation, and positive emotions.Computers and Education: Artificial Intelligence6 (2024), ...

    work page doi:10.1016/j.caeai.2023.100199 2024

  37. [37]

    McGrane, and Sabina Leonelli

    Salvatore Milano, Joshua A. McGrane, and Sabina Leonelli. 2023. Large language models challenge the future of higher education.Nature Machine Intelligence5 (2023), 333–334. https://doi.org/10.1038/s42256-023-00644-2

    work page doi:10.1038/s42256-023-00644-2 2023

  38. [38]

    Susanne Narciss. 2008. Feedback Strategies for Interactive Learning Tasks. InHandbook of Research on Educational Communications and Technology(3rd ed.), J. M. Spector, M. D. Merrill, J. J. G. van Merrienboer, and M. P. Driscoll (Eds.). Taylor & Francis/Routledge, 125–144

    2008

  39. [39]

    Susanne Narciss, Sergey Sosnovsky, Lenka Schnaubert, Eric Andrès, Anja Eichelmann, George Goguadze, and Erica Melis. 2014. Exploring feedback and student characteristics relevant for personalizing feedback strategies.Computers & Education71 (2014), 56–76. https://doi.org/10.1016/j.compedu.2013.09.011

    work page doi:10.1016/j.compedu.2013.09.011 2014

  40. [40]

    Nicol and Debra Macfarlane-Dick

    David J. Nicol and Debra Macfarlane-Dick. 2006. Formative assessment and self-regulated learning: a model and seven principles of good feedback practice.Studies in Higher Education31, 2 (2006), 199–218. https://doi.org/10.1080/ 03075070600572090 AI Assistance for Discretionary Work: Increasing Feedback Provision in Higher Education 23

    2006

  41. [41]

    Shakked Noy and Whitney Zhang. 2023. Experimental evidence on the productivity effects of generative artificial intelligence.Science381, 6654 (2023), 187–192. https://doi.org/10.1126/science.adh2586

    work page doi:10.1126/science.adh2586 2023

  42. [42]

    1988.Organizational citizenship behavior: The good soldier syndrome.Lexington books/DC heath and com

    Dennis W Organ. 1988.Organizational citizenship behavior: The good soldier syndrome.Lexington books/DC heath and com

    1988

  43. [43]

    Kritish Pahi, Shiplu Hawlader, Eric Hicks, Alina Zaman, and Vinhthuy Phan. 2024. Enhancing active learning through collaboration between human teachers and generative AI.Computers and Education Open6 (2024), 100183. https://doi.org/10.1016/j.caeo.2024.100183

    work page doi:10.1016/j.caeo.2024.100183 2024

  44. [44]

    Parasuraman, T.B

    R. Parasuraman, T.B. Sheridan, and C.D. Wickens. 2000. A model for types and levels of human interaction with automation.IEEE Transactions on Systems, Man, and Cybernetics - Part A: Systems and Humans30, 3 (2000), 286–297. https://doi.org/10.1109/3468.844354

    work page doi:10.1109/3468.844354 2000

  45. [45]

    Catherine Paterson, Nathan Paterson, William Jackson, and Fiona Work. 2020. What are students’ needs and preferences for academic feedback in higher education? A systematic review.Nurse Education Today85 (2020), 104236. https: //doi.org/10.1016/j.nedt.2019.104236

    work page doi:10.1016/j.nedt.2019.104236 2020

  46. [46]

    Sida Peng, Eirini Kalliamvakou, Peter Cihon, and Mert Demirer. 2023. The Impact of AI on Developer Productivity: Evidence from GitHub Copilot. arXiv:2302.06590 [cs.SE] https://arxiv.org/abs/2302.06590

    Pith/arXiv arXiv 2023

  47. [47]

    Kjeld Schmidt and Liam Bannon. 1992. Taking CSCW seriously: Supporting articulation work.Computer supported cooperative work (CSCW)1, 1 (1992), 7–40. https://doi.org/10.1007/BF00752449

    work page doi:10.1007/bf00752449 1992

  48. [48]

    2012.Digital Labor: The Internet As Playground and Factory

    Trebor Scholz (Ed.). 2012.Digital Labor: The Internet As Playground and Factory. Taylor & Francis Group

    2012

  49. [49]

    Paschal Sheeran and Thomas L Webb. 2016. The intention–behavior gap.Social and personality psychology compass 10, 9 (2016), 503–518

    2016

  50. [50]

    Ben Shneiderman. 2020. Human-Centered Artificial Intelligence: Reliable, Safe & Trustworthy.arXiv preprint arXiv:2002.04087(2020). https://arxiv.org/abs/2002.04087

    arXiv 2020

  51. [51]

    Valerie J. Shute. 2008. Focus on Formative Feedback.Review of Educational Research78, 1 (2008), 153–189. https: //doi.org/10.3102/0034654307313795

    work page doi:10.3102/0034654307313795 2008

  52. [52]

    Susan Leigh Star and Anselm Strauss. 1999. Layers of silence, arenas of voice: The ecology of visible and invisible work.Computer supported cooperative work (CSCW)8, 1 (1999), 9–30. https://doi.org/10.1023/A:1008651105359

    work page doi:10.1023/a:1008651105359 1999

  53. [53]

    1987.Plans and situated actions: The problem of human-machine communication

    Lucille Alice Suchman. 1987.Plans and situated actions: The problem of human-machine communication. Cambridge university press

    1987

  54. [54]

    Danielle R Thomas, Jionghao Lin, Erin Gatz, Ashish Gurung, Shivang Gupta, Kole Norberg, Stephen E Fancsali, Vincent Aleven, Lee Branstetter, Emma Brunskill, and Kenneth R Koedinger. 2024. Improving Student Learning with Hybrid Human-AI Tutoring: A Three-Study Quasi-Experimental Investigation. InProceedings of the 14th Learning Analytics and Knowledge Conf...

    work page doi:10.1145/3636555.3636896 2024

  55. [55]

    Carlos Toxtli, Siddharth Suri, and Saiph Savage. 2021. Quantifying the Invisible Labor in Crowd Work.Proc. ACM Hum.-Comput. Interact.5, CSCW2, Article 319 (Oct. 2021), 26 pages. https://doi.org/10.1145/3476060

    work page doi:10.1145/3476060 2021

  56. [56]

    Maya Usher. 2025. Generative AI vs. instructor vs. peer assessments: a comparison of grading and feedback in higher education.Assessment & Evaluation in Higher Education50, 6 (2025), 912–927. https://doi.org/10.1080/02602938.2025. 2487495

    work page doi:10.1080/02602938.2025 2025

  57. [57]

    Wang, Ana T

    Rose E. Wang, Ana T. Ribeiro, Carly D. Robinson, Susanna Loeb, and Dora Demszky. 2025. Tutor CoPilot: A Human-AI Approach for Scaling Real-Time Expertise. arXiv:2410.03017 [cs.CL] https://arxiv.org/abs/2410.03017

    arXiv 2025

  58. [58]

    Yu, and Qingsong Wen

    Shen Wang, Tianlong Xu, Hang Li, Chaoli Zhang, Joleen Liang, Jiliang Tang, Philip S. Yu, and Qingsong Wen. 2024. Large Language Models for Education: A Survey and Outlook. arXiv:2403.18105 [cs.CL] https://arxiv.org/abs/2403.18105

    arXiv 2024

  59. [59]

    Lixiang Yan, Lele Sha, Linxuan Zhao, Yuheng Li, Roberto Martinez-Maldonado, Guanliang Chen, Xinyu Li, Yueqiao Jin, and Dragan Gašević. 2024. Practical and ethical challenges of large language models in education: A systematic scoping review.British Journal of Educational Technology55, 1 (2024), 90–112. https://doi.org/10.1111/bjet.13370

    work page doi:10.1111/bjet.13370 2024

  60. [60]

    Qian Yang, Aaron Steinfeld, Carolyn Rosé, and John Zimmerman. 2020. Re-examining Whether, Why, and How Human-AI Interaction Is Uniquely Difficult to Design. InProceedings of the 2020 CHI Conference on Human Factors in Computing Systems(Honolulu, HI, USA)(CHI ’20). Association for Computing Machinery, New York, NY, USA, 1–13. https://doi.org/10.1145/331383...

    work page doi:10.1145/3313831.3376301 2020

  61. [61]

    Jean-Gabriel Young, Amanda Casari, Katie McLaughlin, Milo Z Trujillo, Laurent Hébert-Dufresne, and James P Bagrow. 2021. Which contributions count? Analysis of attribution in open source. In2021 IEEE/ACM 18th International Conference on Mining Software Repositories (MSR). IEEE, 242–253

    2021

  62. [62]

    Padilla, Jeffrey Caterino, Ping Zhang, and Dakuo Wang

    Shao Zhang, Jianing Yu, Xuhai Xu, Changchang Yin, Yuxuan Lu, Bingsheng Yao, Melanie Tory, Lace M. Padilla, Jeffrey Caterino, Ping Zhang, and Dakuo Wang. 2024. Rethinking Human-AI Collaboration in Complex Medical Decision Making: A Case Study in Sepsis Diagnosis. InProceedings of the 2024 CHI Conference on Human Factors in Computing Systems(Honolulu, HI, U...

    work page doi:10.1145/3613904.3642343 2024

  63. [63]

    Alice Li, Andrew Rice, Devon Rifkin, Shawn Simister, Ganesh Sittampalam, and Edward Aftandilian

    Albert Ziegler, Eirini Kalliamvakou, X. Alice Li, Andrew Rice, Devon Rifkin, Shawn Simister, Ganesh Sittampalam, and Edward Aftandilian. 2024. Measuring GitHub Copilot’s Impact on Productivity.Commun. ACM67, 3 (Feb. 2024), 54–63. https://doi.org/10.1145/3633453

    work page doi:10.1145/3633453 2024

  64. [64]

    Chu, Vito Trianni, Ralf H

    Nikolas Zöller, Julian Berger, Irving Lin, Nathan Fu, Jayanth Komarneni, Gioele Barabucci, Kyle Laskowski, Victor Shia, Benjamin Harack, Eugene A. Chu, Vito Trianni, Ralf H. J. M. Kurvers, and Stefan M. Herzog. 2025. Human–AI collectives most accurately diagnose clinical vignettes.Proceedings of the National Academy of Sciences122, 24 (2025), e2426153122....

    work page doi:10.1073/pnas.2426153122 2025