Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 26 of the 26 references it cites were found in a registry or an index; 7 of the 37 sources its claims cite were checked for support.

Using ChatGPT and Other Large Language Model (LLM) Applications for Academic Paper Assignments

Open access

Abstract

Large language models (LLMs), like ChatGPT, GitHub Copilot, and Microsoft Copilot, present challenges in university education, particularly for paper assignments. These AI-driven tools enable students to (semi)automatically complete tasks that were previously considered evidence of skill acquisition, potentially affecting grading and skill development. However, the use of these tools is not legally considered plagiarism and is becoming increasingly integrated into various software solutions.

Overview

CoChat Summary

TL;DR

This working paper examines how ChatGPT/GPT-4 and other LLMs can support or undermine academic paper assignments, using digital misinformation and German elections as a running example. It argues LLMs are useful for exploration, structuring, hypotheses, and research questions but remain unreliable for case-specific, operationalization, data-analysis, and source-verification tasks.

Key findings

  • ChatGPT/GPT-4 helped generate overviews, potential hypotheses, research questions, and theory lists; for example, it produced five testable hypotheses and linked them to theories such as issue voting, political trust, agenda-setting, political efficacy, and contextual voting.
  • Performance declined as requests became more specific: for the German Federal Election 2021 and 2017, outputs remained general and not specific to the German context; GPT-4’s knowledge cutoff was September 2021, limiting analysis of the 2021 election.
  • References can be plausible but not support the cited claims: in one response all listed sources existed, but Bordino et al. (2012) and Vosoughi et al. (2018) did not cover the economic consequences of misinformation as cited.
  • LLM use is not legally considered plagiarism and is becoming integrated into software; the paper recommends responsible use, disclosure of AI use, and grading emphasis on tasks where models perform poorly.

Methods

  • Conceptual/working paper offering guidelines for students at the Chair for the Governance of Complex and Innovative Technological Systems, University of Bamberg; no empirical sample or systematic benchmark is reported.
  • The author ran prompts through the ChatGPT web interface using the GPT-4 model on March 18 and 19, 2023, with digital misinformation and German Federal Elections 2017/2021 as the running example; responses are described as non-deterministic and path-dependent.
  • Analysis is a qualitative evaluation of model outputs across academic paper tasks: investigating literature, linking a case to a broader category, generating theory-driven hypotheses, formulating research questions, operationalizing expectations, conducting data analysis, and structuring/writing the paper.

Limitations

  • Stated: the notes apply specifically to students at the Chair at the University of Bamberg, and other educators may have different guidelines; prompts and responses are non-deterministic and may not replicate.
  • Stated: the paper focuses on academic paper assignments rather than broader LLM issues, and it uses one model/version and one example topic.
  • Evident: there is no systematic coding, inter-rater reliability, or controlled comparison; conclusions rest on the author’s illustrative prompt sessions and may not generalize to other models, disciplines, or assignments.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.31235/osf.io/d84q6

Cites

26 of 26 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

6 Valid · 27 Not valid · 15 Not checked · 117 Own results and interpretations

12 more claims

Create a free account to see all 165 claims and every source verdict.

See all claims — free account
  • Probably not a substantial claim. 146
  • It may state two separate claims as one. 49

Create a free account to see these claims.

Citations

Cites

Show all 20 references

Research works printed but not linked: 2

  1. Becker, H. S. (2020). Writing for social scientists: How to start and finish your thesis, book, or article (3rd ed.). University of Chicago Press.
  2. Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017). Attention is all you need. In I. Guyon, U. von Luxburg, S. Bengio, H. Wallach, R. Fergus, S. V. N. Vishwanathan, & R. Garnett (Eds.), Nips 2017: 31st conference on neural information processing systems (pp. 1-11, Vol. 30). Curran Associates, Inc. https : / / proceedings . neurips . cc / paper / 2017 / file / 3f5ee243547dee91fbd053c1c4a845aa-Paper.pdf

Cited by

Metadata

Authors
Jungherr, A.
Affiliations
University of Bamberg (DE)
Corresponding author
Jungherr, Andreas
Published
2023-03-24
OpenAlex
W4360939370
Cited by
37
References
20
Citations by year
2026: 2 · 2025: 8 · 2024: 19 · 2023: 8
Type
posted content
Language
en
Open access
Locations
  • SocArXiv (OSF Preprints) (submittedVersion)
  • SocArXiv (OSF Preprints) (submittedVersion, cc-by-nc-sa)
Topics
  • Social Sciences › Education › Artificial Intelligence in Education
  • Social Sciences › Education › Academic integrity and plagiarism
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
Keywords
large language models, ChatGPT, academic writing, research paper assignment, responsible AI use, student plagiarism, academic integrity, social studies education

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-04, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-04 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Using ChatGPT and Other Large Language Model (LLM) Applications for Academic Paper Assignments. https://app.cochat.ai/papers/10.31235/osf.io/d84q6
The paper
Jungherr, A. (2023). Using ChatGPT and Other Large Language Model (LLM) Applications for Academic Paper Assignments. https://doi.org/10.31235/osf.io/d84q6

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.