CoChat
Sign Up
Cited by 21Artificial Intelligence in EducationAcademic integrity and plagiarismArtificial Intelligence in Healthcare and Education

Widen the debate: What is the academic community’s perception on ChatGPT?

Open access
Jiang, Yingying, Xie, Lindai, Lin, Guohui, Mo, Fangfang·2024·Education and Information Technologies
DOI 10.1007/s10639-024-12677-0
Open this paper in CoChat — free accountOpen PDF

Abstract

ChatGPT has surprised academia with its remarkable abilities but also raised substantial concerns regarding academic integrity and misconduct. Despite the debate, empirical research exploring the issue is limited. The purpose of this study is to bridge this gap by analyzing Twitter data to understand how academia is perceiving ChatGPT. A total of 9733 tweets were collected through Python via Twitter API in three consecutive weeks in May and June 2023; and 3000 most relevant ones were analyzed in Atlas ti. 23. Our findings reveal a generally supportive attitude towards using ChatGPT in academia, but the absence of clear policies and regulations requires attention. Discussions primarily focus on academic integrity, learning effectiveness, and teaching efficiency. Tweets from influencers with over one million followers were analyzed separately. The significance of these findings and the limitations of the study are included.

AI snapshot

Written from the paper’s full text.

TL;DR

This study analyzes Twitter data to examine how the academic community perceived ChatGPT in May–June 2023, finding generally supportive attitudes alongside concerns about ethics, equity, and missing policies. It reports that supportive/debatable tweets outnumbered opposing ones by 4.5 to 1, while policy discussion was nearly absent.

Key findings

  • Among 3,000 analyzed tweets, attitudes were 191 supportive, 191 debatable, and 42 against ChatGPT in academia—4.5 times more tweets leaning toward support or discussion than opposition.
  • Ethics was coded in 152 tweets, with 31 specifically about plagiarism; policy discussion was minimal, with only 7 of 3,000 tweets addressing it.
  • Learning tweets were mostly positive (72) versus negative (11); teaching-related tweets (232 total) were mostly positive (30) or debatable (20) versus negative (3); research-related tweets numbered 272, often citing literature-review assistance.
  • Among 65 tweets from influencers with over one million followers, the majority remained positive, emphasizing productivity, plugins, teaching/learning impacts, and responsible AI use.

Methods

  • Collected 9,733 English tweets via Python and the Twitter API (Basic Access Tier) in three consecutive weeks: May 25, June 1, and June 8, 2023; retweets, quotations, replies, Chinese “账号” advertisements, and sports-ranking tweets were excluded.
  • Analyzed the first 1,000 most relevant tweets from each week (3,000 total) in Atlas.ti 23 using a codebook developed from search keywords; three researchers coded the same 100 tweets and achieved 96% inter-rater reliability.
  • Codes covered perspectives (teachers, students, faculty, researchers, other), learning, teaching, research, ethics, equity, policy, and attitude (support, against, debatable).

Limitations

  • The paper states a bias toward “tech-savvy” academics represented in Twitter data and a three-week data collection period that may not fully capture changing or evolving attitudes.
  • The paper also notes API constraints: researchers could extract only the most recent 7-day data because the Academic Research tier was unavailable, and raw data sharing is constrained by Twitter’s terms of service.
  • Evident from the text: the sample is English-only, filtered by selected keywords and exclusions, and drawn from Twitter users rather than a verified academic population, which may limit generalizability.

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

Integrity signals

Registry

Crossref · 10.1007/s10639-024-12677-0

Cites

64 of 65 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 32 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    Out of 3,000 analyzed tweets, 152 were coded as related to 'Ethics'.

    “Out of 3,000 analyzed tweets, 152 were coded as related to”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.
  • Findingown data

    The academic community's attitude towards using ChatGPT in academia is generally supportive.

    “Our findings reveal a generally supportive attitude towards using ChatGPT in academia, but the absence of clear policies and regulations requires attention.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.
  • Findingown data

    Discussion around policy relating to the use of ChatGPT was minimal in the dataset, with only seven out of 3,000 tweets addressing this topic.

    “Discussion around policy relating to the use of ChatGPT was minimal in the dataset, with only seven out of 3,000 tweets addressing this topic.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.
  • Findingown data

    Of the 152 ethics-related tweets, 31 were specifically about plagiarism.

    “Of these, 31 were specifically about plagiarism.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.
  • Findingown data

    Among the 3000 analyzed tweets, recorded attitudes towards using ChatGPT for learning are mostly positive (72) rather than negative (11).

    “Among the 3000 analyzed tweets, recorded attitudes towards using ChatGPT for learning are mostly positive (72) rather than negative (11).”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.
  • Classified as probably not a substantial claim. 36
  • May combine two claims. 27
  • The quote only partly states this claim. 1

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    ChatGPT's strengths and weaknesses were shown in its application in STEM learning; users shared it was a great help in learning computer science subjects.

    “ChatGPT's strengths and weaknesses were shown in its application in STEM learning.Users shared it was a great help in learning computer science subjects,”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingreasoning

    Many tweets highlighted ChatGPT's usefulness for untraditional college students and lower-income individuals who may not afford a college education, therefore fostering educational equity.

    “Many tweets highlighted the ChatGPT's usefulness for untraditional college students and lower-income individuals who may not afford a college education, therefore fostering educational equity.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given

    Some individuals reported changes in their assessment formats, such as a shift from computer exams to paper-and-pencil tests.

    “Indeed, some individuals reported changes in their assessment formats, such as a shift from computer exams to paper-and-pencil tests.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Some users expressed a progressive perspective, seeing the plagiarism controversy as an opportunity for change.

    “However, some users expressed a more progressive perspective, seeing this controversy as an opportunity for change.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • Classified as probably not a substantial claim.

    Findingreasoning

    Users generally reached a consensus that ChatGPT's answers/output need to be manually rechecked and revised for effective learning and high-quality task completion.

    “To summarize while affirming the ability and helpfulness of ChatGPT, users generally reached a consensus that (1) it is necessary to provide ChatGPT with sufficient information and accurate questions, and (2) the answers/output of ChatGPT needs to be manually rechecked and revised for the purpose of effective learning and complete tasks with high-quality.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • Classified as probably not a substantial claim.

    Findingreasoning

    Users generally reached a consensus that it is necessary to provide ChatGPT with sufficient information and accurate questions.

    “To summarize while affirming the ability and helpfulness of ChatGPT, users generally reached a consensus that (1) it is necessary to provide ChatGPT with sufficient information and accurate questions, and (2) the answers/output of ChatGPT needs to be manually rechecked and revised for the purpose of effective learning and complete tasks with high-quality.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    Popular opinions include that ChatGPT helps save time and the process is similar to interacting with a knowledgeable person.

    “Popular opinions include that it helps save time and the process is similar to interacting with a knowledgeable person.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing).

    Findingnone given

    Another user pointed out ChatGPT's weakness in simple statistics.

    “On the other hand, another user pointed out its weakness in simple statistics”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingnone given

    Concerns were raised about AI's unintentional bias, particularly towards non-English speakers, and the potential gender bias inherent in language.

    “However, concerns were raised about AI's unintentional bias, particularly towards non-English speakers, and the potential gender bias inherent in language”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • May combine two claims. The quote only partly states this claim.

    Findingown data

    Among tweets discussing the use of ChatGPT to learn, 'prompts' is frequently mentioned, and users emphasized the importance of using appropriate prompts to maximize ChatGPT's effectiveness.

    “Among the tweets discussing the use of Chat-GPT to learn,”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • May combine two claims.

    Findingown data

    Strong opinions were formed around the plagiarism issue, with tweets describing ChatGPT as a 'plagiarism machine'.

    “Strong opinions were formed around the plagiarism issue, with tweets describing ChatGPT as a”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Educators shared frustration over students' dependence on ChatGPT.

    “Educators shared frustration over students' dependence on ChatGPT”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingcitation

    Our findings align with Kostka and Toncelli (2023) in highlighting the potential positive impact of integrating AI, particularly ChatGPT, in teaching.

    “Our findings, align with the study of Kostka and Toncelli (2023), highlighting the potential positive impact of integrating AI, particularly ChatGPT in teaching.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Kostka and Toncelli, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable).

    Findingown data

    Our findings emphasize the crucial need for corresponding training and professional development for teachers.

    “Furthermore, our findings emphasize the crucial need for corresponding training and professional development for teachers.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingreasoning

    Tweets emphasized the need for training programs to empower educators with knowledge and skills for effective integration of AI tools such as ChatGPT.

    “Teacher professional development and learning emphasize the crucial need for training programs that empower educators with the essential knowledge and skills required for the effective integration of AI tools, such as ChatGPT within the classroom.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • May combine two claims.

    Findingown data

    Top tweets praised ChatGPT's productivity enhancements and valuable plugins, highlighting the transformative impact of AI on human interactions and cognition.

    “Top tweets praised ChatGPT's productivity enhancements and valuable plugins, highlighting the transformative impact of AI on human interactions and cognition.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    The overall sentiment about the use of ChatGPT in teaching-related activities reflected a majority of positive opinions (30 tweets) and debatable perspectives (20 tweets) rather than negative (3 tweets).

    “The overall sentiment reflected a majority of positive opinions (30 tweets) and debatable perspectives (20 tweets) rather than negative (3 tweets) about the use of ChatGPT in teaching-related activities.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Concerns surrounding the need and ethical considerations to adapt teaching practices in an AI-driven world are raised.

    “Furthermore, concerns surrounding the need and ethical considerations to adapt teaching practices in an AI-driven world are also raised.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • May combine two claims.

    Findingown data

    Top tweets mentioned the importance of responsible AI use and ongoing research to effectively address ethical challenges.

    “Lastly, the importance of responsible AI use and ongoing research to effectively address ethical challenges is mentioned,”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Findingown data

    Our results reveal that the primary discussion topics included concerns related to ethics, especially plagiarism, and hindered equity issues with AI.

    “Our Results reveal the primary discussion topics included concerns related to ethics, especially plagiarism, and hindered equity issues with AI.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim. May combine two claims.

    Findingcitation

    The range of benefits in our study suggests personalized learning, promotion of critical thinking, encouragement of exploration and creativity, support for inquiry and problem-solving, and enhancement of students' understanding of real-world connections.

    “The range of benefits in our study suggest personalized learning, promotion of critical thinking, encouragement of exploration and creativity, support for inquiry and problem-solving, and the enhancement of students' understanding of real-world connections (Exintaris et al., 2023;Limo et al., 2023;Morath et al., 2023).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Exintaris et al., 2023 source

      Not checked in the free run.

    • cites Limo et al., 2023

      Not checked in the free run.

    • cites Morath et al., 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    According to tweets, ChatGPT demonstrated proficiency in handling 'a range of monotonous jobs,' showcasing its versatility in tasks such as brainstorming ideas or drafting an outline for a research paper.

    “According to tweets, ChatGPT demonstrated proficiency in handling”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingreasoning

    The sentiment echoes positivity, indicating widespread interest and consideration of the applicability of AI tools across various educational domains, from STEM to art education.

    “The sentiment echoes this positivity, indicating widespread interest and consideration of the applicability of AI tools across various educational domains, from STEM to art education.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingown data

    Educators reflected on the ethical aspects of incorporating ChatGPT into their teaching approaches and how to effectively communicate these considerations to their students.

    “Educators reflected on the ethical aspects of incorporating ChatGPT into their teaching approaches and how to effectively communicate these considerations to their students”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • May combine two claims.

    Findingown data

    There were discussions about misuse of ChatGPT and misuse of the AI-generated plagiarism machine to detect bot-processed languages.

    “We see these issues with teachers expressing frustration with students' dependence on ChatGPT on course assignments, and discussions about the misuse of ChatGPT, also the misuse of the AI-generated plagiarism machine to detect bot-processed languages.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingcitation

    Our results align with previous findings that AI can enhance the learning experience when used responsibly and appropriately.

    “Our results align with previous findings that AI can enhance the learning experience when used responsibly and appropriately (e.g., Baidoo-Anu & Owusu Ansah, 2023;Javaid et al., 2023).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Baidoo-Anu & Owusu Ansah, 2023

      Not checked in the free run.

    • cites Javaid et al., 2023 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Findingnone given

    Educators discuss how AI technologies can enhance student engagement and provide new learning opportunities.

    “Educators also discuss how AI technologies can enhance student engagement and provide new learning opportunities.”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • May combine two claims.

    Findingown data

    Top tweets showcased diverse applications of ChatGPT beyond education, including enhancing social skills, providing job interview support, and assisting in crypto research.

    “Secondly, these tweets showcased diverse applications of ChatGPT beyond education, including enhancing social skills, providing job interview support, and assisting in crypto research,”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Some tweets encourage learners to critically assess responses from ChatGPT.

    “Additionally, some tweets encourage learners to critically assess the responses from ChatGPT.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    Users reported that ChatGPT proved helpful in summarizing important articles.

    “Users reported that the tool proved helpful in summarizing important articles, with one tweet expressing,”
    Quote in the paperNot a substantial claim
    • Rests on an assumption: Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.
    • Rests on an assumption: The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.
    • Rests on an assumption: The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

  • May combine two claims.

    Findingown data

    Twitter users' experiences highlight ChatGPT's capabilities in brainstorming ideas, drafting research outlines, and summarizing critical articles for literature reviews.

    “Tweeter users' experiences highlight the AI's capabilities in brainstorming ideas, drafting research outlines, and, notably, summarizing critical articles for literature reviews.”
    Quote in the paperSubstantial claim

  • May combine two claims.

    Findingown data

    The study reveals concerns regarding ethics, discrimination, and absence of regulations.

    “However, it also reveals concerns regarding ethics, discrimination, and absence of regulations.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingcitation

    Educators on Twitter acknowledged the importance of acquiring the necessary knowledge and skills for the effective integration of AI tools.

    “Educators on Twitter acknowledged the importance of acquiring the necessary knowledge and skills for the effective integration of AI tools (Malik et al., 2023) .”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Malik et al., 2023 source

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    The majority of existing work on ChatGPT in academia consists of conceptual and opinion-based discussions and quantitative surveys aimed at specific groups.

    “The majority of existing work consists of conceptual and opinion-based discussions (e.g., Farrelly & Baker, 2023;Hsu & Ching., 2023;Sohail et al., 2023;Yu, 2023), along with quantitative surveys aimed at specific groups (e.g., Hammour et al., 2023;Tangadulrat et al., 2023).”
    Quote in the paperSubstantial claim0 of 3 sources checked
    • cites Farrelly & Baker, 2023 source

      Not checked in the free run.

    • cites Hsu & Ching., 2023 source

      Not checked in the free run.

    • cites Sohail et al., 2023 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Rospigliosi's (2023) perspective is that when interacting with ChatGPT, learners should assess, integrate and comprehend.

    “This aligns with Rospigliosi's (2023) perspective that when interacting with ChatGPT, learners should assess, integrate and comprehend.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Rospigliosi, 2023 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    These issues (ethics, plagiarism, and equity concerns) exist largely due to the absence of government-specific policies on AI and university-level regulations for incorporating AI in class and research.

    “However, these issues exist largely due to the absence of government-specific policies on the use of AI, as well as the absence of universitylevel regulations for incorporating AI in class and in research.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    There are concerns about ethical implications, misuse, and the potential impact on different domains as the AI community celebrates progress in natural language generation.

    “However, as the AI community celebrates the progress in natural language generation, there are also concerns about ethical implications, misuse, and the potential impact on different domains (Ray, 2023;Zhou et al., 2023).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Ray, 2023

      Not checked in the free run.

    • cites Zhou et al., 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Historically, AI has been utilized in education, such as computerized adaptive testing and word-processing auto-corrections.

    “Historically, AI has been utilized in Education like computerized adaptive testing and word-processing auto-corrections (Czerkawski et al., 2023;Mujtaba & Mahapatra, 2020).”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Czerkawski et al., 2023

      Not checked in the free run.

    • cites Mujtaba & Mahapatra, 2020

      Not checked in the free run.

  • May combine two claims.

    Study design and conductown data

    The same 100 tweets were coded by three researchers and achieved inter-rater reliability of 96%.

    “The same 100 tweets were coded by three researchers and achieved inter-rater reliability of 96%.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductcitation

    The researchers were restricted to extracting the most recent 7-day data from Recent Search Endpoint due to the unavailability of the Academic Research tier at the time of the study.

    “The researchers were restricted to extracting the most recent 7-day data from Recent Search Endpoint due to the unavailability of the Academic Research tier at the time of the study (Twitter Community, 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Twitter Community, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    There has been a noticeable shift in education's priorities away from solely achievement scores toward understanding students' learning journeys and their overall wellbeing.

    “Furthermore, there's been a noticeable shift in education's priorities these days.Rather than placing emphasis solely on achievement scores, there's a growing focus on understanding students' learning journeys and their overall wellbeing (Jie et al., 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Jie et al., 2023

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    There is an absence and inconsistency in policies and regulations on using ChatGPT.

    “The absence and inconsistency in policies and regulations on using ChatGPT require a broader debate and empirical research into this tool and similar AI tools.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites College Rover, 2023

      Not checked in the free run.

  • May combine two claims.

    Premisecitation

    A survey of 1,000 college students revealed that approximately 70% of college students use ChatGPT multiple times a month, and 41% use it weekly.

    “A survey of 1,000 college students revealed that approximately 70% of college students use ChatGPT multiple times a month, and 41% use it weekly.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites College Rover, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    Duplicate records were identified and removed to ensure data integrity and eliminate redundancy.

    “Duplicate records were identified and removed to ensure data integrity and eliminate redundancy.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Interpretationown data

    Precise and thoughtful prompting is essential for acquiring tailored and accurate answers from ChatGPT.

    “However, we also noticed a critical aspect of effective ChatGPT use for learning: the art of prompt asking.Precise and thoughtful prompting is essential for acquiring tailored and accurate answers-this requires learners to ask accurate questions and provide relevant background information.”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Interpretationown data

    Our research methodology, utilizing Twitter as a lens to examine academic discourse, demonstrates the value of social media data in revealing academia's reaction to emerging technologies.

    “In addition, our research methodology, utilizing Twitter as a lens to examine academic discourse, demonstrates the value of social media data in revealing academia's reaction to emerging technologies.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Previous literature suggests challenges include the potential for encountering fabricated information and outdated domain knowledge, concerns about decision-making accountability, and the opportunity cost of excessive reliance on ChatGPT.

    “However, as previous literature suggests challenges include the potential for encountering fabricated information and outdated domain knowledge alongside concerns about decision-making accountability and the opportunity cost of excessive reliance on ChatGPT (Walters & Wilder, 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Walters & Wilder, 2023 source

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    This approach enriches understanding of current attitudes and engagement towards AI in academia and demonstrates the utility of social media analytics as a dynamic tool for rapid and exploratory research.

    “This approach not only enriches our understanding of current attitudes and engagement towards AI in academia but also demonstrates the utility of social media analytics as a dynamic tool for conducting rapid and exploratory research.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductnone given

    The three-week data collection period may not fully capture changing/evolving attitudes.

    “Some notable limitations include a bias towards”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    ChatGPT has revealed benefits in improving writing skills and identifying key themes.

    “ChatGPT, as a conversational AI, has fundamentally altered how information is retrieved (Wei et al., 2023), and revealed its benefits in improving writing skills and identifying key themes (Shidiq, 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Shidiq, 2023

      Not checked in the free run.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Over the past year, diverse research fields have explored ChatGPT, including environmental research, physics education, psychiatry, and healthcare education.

    “We acknowledge that over the past year, diverse research fields have explored Chat-GPT including Environmental Research (Zhu et al., 2023), Physics Education (Kieser et al., 2023); Psychiatry (Cheng et al., 2023); Healthcare Education (Sallam, 2023) and more.”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Zhu et al., 2023

      Not checked in the free run.

    • cites Kieser et al., 2023 source

      Not checked in the free run.

    • cites Cheng et al., 2023

      Not checked in the free run.

51 of 51

  • The paper relies on this without support.

    Twitter data effectively captures real-time discussions and sentiments within the academic community regarding AI technologies.

    “Twitter is a publicly accessible platform which is effective in capturing real-time discussions and sentiments within the academic community regarding AI technologies.”

  • The paper relies on this without support.

    The keywords developed from the literature review ensure a comprehensive and focused representation of academic discourse with ChatGPT.

    “This strategy ensured the comprehensive and focused representation of academic discourse with ChatGPT in this study.”

  • The paper relies on this without support.

    The 3000 tweets analyzed (the first 1000 most relevant tweets from three consecutive weeks) are representative of the broader academic discourse on ChatGPT.

    “For this study, researchers analyzed 3000 tweets (the first 1000 most relevant tweets from three consecutive weeks) in Atlas.ti23.”

  • The paper relies on this without support.

    The codebook developed based on search keywords adequately captures the perceptions and uses of ChatGPT in academia.

    “A codebook was developed based on search keywords (Table 1).”

  • The paper relies on this without support.

    Many university-level policies on AI in education are currently in development and available for internal consultation.

    “It's worth noting that many university-level policies are currently in development and are available for internal consultation, indicating a move towards more structured and guided use of AI in educational contexts.”

  • The paper relies on this without support.

    AI is here to stay.

    “As many experts said, AI is here to stay.”
6 of 6

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Discussions surrounding ChatGPT are highly concentrated in educational and normative domains, creating a degree of path dependency.

    “Discussions surrounding ChatGPT are highly concentrated in educational and normative domains, creating a degree of path dependency (Jiang et al., 2024).”

    From The Attention Mismatch: Mapping the Structural Academic Governance Deficit in the Age of Generative AI (2026) · cites this paper as Jiang et al., 2024

    Sections 6.4 (Teaching) and 6.2.1 (Ethics) document discussion of educational uses and ethical concerns in the academic tweets studied. But the paper does not establish that ChatGPT discussions generally are highly concentrated in those domains or that this creates path dependency; its findings concern a search specifically focused on academic discourse (Keywords section). (LLM, conf=0.82)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Learning support count differs between results and table

    Major

    The results give 72 positive learning-related tweets, while Table 2 lists 81 supportive tweets for learning. Both give 11 negative or against tweets for that category.

  • Combined supportive and debatable tweets are not 4.5 times opposing tweets

    Minor

    The stated ratio gives 191 supportive and 191 debate-inviting tweets. Together, those are 382 tweets-about 9.1 times the 42 opposing tweets. The 4.5-fold figure applies to either group separately, not to both together.

2 of 2

Cites

Flagged first, then by citations

Research works printed but not linked: 5

  1. Czerkawski, B., Brylska, K., & Gackowski, T. (2023). July). Using AI in Scholarly Research: Advantages and drawbacks. EdMedia + Innovate Learning (pp. 1482–1488). Association for the Advancement of Computing in Education (AACE).
  2. Fitria, T. N. (2023). March). Artificial intelligence (AI) technology in OpenAI ChatGPT application: A review of ChatGPT in writing English essay. ELT Forum: Journal of English Language Teaching (Vol, 12(1), 44–58.
  3. Limo, F. A. F., Tiza, D. R. H., Roque, M. M., Herrera, E. E., Murillo, J. P. M., Huallpa, J. J., & Gonzáles, J. L. A. (2023). Personalized tutoring: ChatGPT as a virtual tutor for personalized learning experiences. Social Space, 23(1), 293–312.
  4. Shidiq, M. (2023, May). The use of artificial intelligence-based chat-gpt and its challenges for the world of education; from the viewpoint of the development of creative writing skills. In Proceeding of International Conference on Education, Society and Humanity (Vol. 1, No. 1, pp. 353–357).
  5. Sun, G. H., & Hoelscher, S. H. (2023). The ChatGPT storm and what faculty can do. Nurse Educator, 48(3), 119–124.

Cited by

Sorted by citations

Authors
Jiang, Y., Xie, L., Lin, G., & Mo, F.
Affiliations
  • Purdue University West Lafayette (US)
  • Virginia Commonwealth University (US)
Corresponding author
Jiang, Yingying
ORCID
Venue
Education and Information Technologies
Publisher
Springer Science and Business Media LLC
Published
2024-04-13
Volume
29
Issue
15
Pages
20181-20200
ISSN
  • 1360-2357 (ISSN-L)
  • 1573-7608
DOI
10.1007/s10639-024-12677-0
OpenAlex
W4394783870
Semantic Scholar
95de4949d333eda4f35a23a91e467ddf6541a2b1
Cited by
21
References
65
Citations by year
2026: 3 · 2025: 14 · 2024: 4
FWCI
7.91
Type
journal article
Language
en
Open access
Licence
Locations
  • Education and Information Technologies (publishedVersion, cc-by)
  • Purdue e-Pubs (Purdue University System) (submittedVersion, cc-by)
Topics
  • Physical Sciences › Computer Science › Information Systems › Artificial Intelligence in Education
  • Social Sciences › Safety Research › Academic integrity and plagiarism
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
Keywords
ChatGPT, academic community perceptions, academic integrity, academic misconduct, generative AI in education, Twitter discourse, AI policy and regulation, learning effectiveness, teaching efficiency, social media influencers

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.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-06.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-06.

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