Cited by 197Artificial Intelligence in EducationArtificial Intelligence in Healthcare and EducationEducational Innovations and Challenges

Challenges for Higher Education in the Era of Widespread Access to Generative AI

Open access
Walczak, Krzysztof, Cellary, Wojciech·2023·Economics and Business Review

Abstract

The aim of this paper is to discuss the role and impact of generative artificial intelligence (AI) systems in higher education. The proliferation of AI models such as GPT-4, Open Assistant and DALL-E presents a paradigm shift in information acquisition and learning. This transformation poses substantial challenges for traditional teaching approaches and the role of educators. The paper explores the advantages and potential threats of using generative AI in education and necessary changes in curricula. It further discusses the need to foster digital literacy and the ethical use of AI. The paper’s findings are based on a survey conducted among university students exploring their usage and perception of these AI systems. Finally, recommendations for the use of AI in higher education are offered, which emphasize the need to harness AI's potential while mitigating its risks. This discourse aims at stimulating policy and strategy development to ensure relevant and effective education in the rapidly evolving digital landscape.

AI snapshot

Written from the paper’s full text.

TL;DR

Walczak surveys university students on their use and perception of generative AI and argues that widespread access to tools like GPT-4 and DALL-E disrupts higher education's traditional role in knowledge transfer and skill development. The paper recommends that universities teach AI literacy, critical verification of AI output, and preserve knowledge-based examination while preparing students for lifelong self-learning.

Key findings

  • Of ~1,000 students contacted, 143 responded; 81% reported prior use of online AI systems and 79% had used ChatGPT. The most common purpose was "out of curiosity" (88%), followed by "finding a solution to a problem" (63%), "information search" (62%) and "programming" (52%); "generating an essay/abstract" and "translation" were less frequent.
  • Only 2% of students declared complete trust in AI-generated content; 24% said they always scrutinize AI-produced data and facts, and 73% verify via search engines (27% via Wikipedia). Yet when shown a GPT-3.5 biography of Adam Mickiewicz containing a factual error, only 21% identified it as false while 50.7% failed to do so; among students who claimed to always verify, 45.7% still missed the error.
  • 31% of students thought AI systems are currently mature enough for educational integration, but 76% agreed on their future utility; 55% advocated integrating AI into education with instruction on its use, and 51% were reluctant to have exams graded by AI.
  • Better-performing students appeared to use AI less: in the "very good" cohort, 58% declared no use of AI in education (including 20% who tried and stopped). 89% of students knew AI does not think like a human.

Methods

  • A survey of university students from diverse institutions and disciplines (Computer Science, Econometrics, Cloud Solutions, IoT, E-Business, Industry 4.0, Management, Administration, Business Psychology, HRM); 143 of about 1,000 responded. 36% were bachelor's students, 64% master's; 43% female, 57% male.
  • The survey covered chatbot interaction preferences, purposes of ChatGPT use, other AI systems used, frequency of AI use for academic tasks, opinions on AI in education, trust in AI content, and a test in which students judged a GPT-3.5-generated biography containing a deliberate factual error.
  • Results are reported as descriptive percentages and cross-tabulations (e.g., declared trust vs. actual error detection; AI use frequency vs. self-assessed performance). The paper also reviews generative AI architectures (VAEs, GANs, Transformers, GPT, WaveNet) and discusses hallucinations, authorship and originality.

Limitations

  • The paper itself notes that the sample of students declaring very poor performance was too small to draw statistically significant conclusions, and that it is premature to draw definitive conclusions about AI's future trajectory.
  • Evident from the text: the survey response rate is low (143 of ~1,000, ~14%), the sample is concentrated in computer-science-adjacent fields at the author's institution, and the analysis is largely descriptive with no inferential statistics reported.
  • Evident from the text: the paper's own example of GPT hallucination concerns the author's personal biography, and the recommendations are framed as preliminary rather than tested interventions.

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

Integrity signals

Registry

Crossref · 10.18559/ebr.2023.2.743

Cites

34 of 34 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

0 of 11 cited sources checked · 0 supported · 0 flagged

  • Findingown data

    1% of students believe that AI emulates human cognition.

    “and a mere 1% believe that AI emulates human cognition.”
    Quote in the paperNot a substantial claim
  • Findingown data

    22% of respondents reported using GPT during class sessions.

    “However, the use of GPT during class sessions appears to be less frequent as it was reported by only 22% of respondents.”
    Quote in the paperNot a substantial claim
  • Findingown data

    89% of students are aware that AI does not think like a human.

    “In terms of understanding the cognitive function of AI 89% of students are aware that AI does not think like a human”
    Quote in the paperNot a substantial claim
  • Findingown data

    20% of students are in favour of permitted use of AI systems but without active promotion.

    “Conversely 20% of students are in favour of their permitted use but without active promotion.”
    Quote in the paperNot a substantial claim
  • Findingown data

    Only 2% of students declared complete trust in content generated by AI systems.

    “Only 2% of students declared complete trust in the content generated by AI systems (Figure 9, left).”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim. 73
  • May combine two claims. 26
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    20% of students remained neutral on whether AI systems are mature enough for integration in the educational process.

    “20% remained neutral.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    10% of students are unsure whether AI thinks like a human.

    “10% are unsure”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given

    65% of students are not worried about job displacement due to AI.

    “Interestingly 65% of students are not worried about job displacement due to AI”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingnone given

    86% of students envisage that AI systems will serve professional functions in the future.

    “This belief is amplified when considering future scenarios with a substantial 86% of students envisaging that AI systems will serve professional functions in the future.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Findingown data

    38% of students consider that AI systems have reached a sufficient level of maturity for professional applications in specialized domains.

    “A total of 38% of students consider that AI systems have reached a sufficient level of maturity for professional applications albeit in specialized domains.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    87.5% of those who use AI often or very often believe the use of AI in education should be allowed and students should be encouraged and taught how to use AI.

    “87.5% who use AI often or very often”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    55.24% of all participants believe the use of AI in education should be allowed and students should be encouraged and taught how to use AI.

    “This opinion has been expressed by 55.24% of all participants”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Findingown data

    61.9% of those who ever tried using AI believe the use of AI in education should be allowed and students should be encouraged and taught how to use AI.

    “61.9% of those who ever tried using AI”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing), S4 (based).

    Findingnone given

    In the 'very good' cohort of students, 58% declared no use of AI in the education process.

    “In the”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Hallucinations can be categorized broadly into closed-domain and open-domain hallucinations.

    “Hallucinations can be categorized broadly into closed-domain and opendomain hallucinations (Bubeck et al., 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Bubeck et al., 2023

      Not checked in the free run.

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

    Premisecitation

    Large Language Models (LLMs) have emerged as the dominant paradigm in text processing and generation, outperforming traditional methods across a wide range of NLP tasks.

    “In recent years Large Language Models (LLMs) have emerged as the dominant paradigm in the realm of text processing and generation, outperform-ing traditional methods and demonstrating unprecedented performance across a wide range of natural language processing tasks”
    Quote in the paperNot a substantial claim0 of 3 sources checked
    • cites Brown et al., 2020

      Not checked in the free run.

    • cites Devlin et al., 2019

      Not checked in the free run.

    • cites Vaswani et al., 2017

      Not checked in the free run.

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

    Premisecitation

    AI systems can generate realistically sounding voices or music.

    “The model is able to generate realistically sounding voices or music by directly modelling waveforms with the use of a neural network trained with sample recordings”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites van den Oord et al., 2016

      Not checked in the free run.

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

    Premisecitation

    AI image generators can synthesize high-quality, realistic images.

    “These generators employ deep learning techniques, specifically Variational Autoencoders (VAEs) (Kingma & Welling, 2014), Generative Adversarial Networks (GANs) (Goodfellow et al., 2014), and similar frameworks to synthesize high-quality, realistic images.”
    Quote in the paperNot a substantial claim0 of 2 sources checked
    • cites Kingma & Welling, 2014

      Not checked in the free run.

    • cites Goodfellow et al., 2014

      Not checked in the free run.

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

    Study design and conductown data

    A total of 143 students out of about 1,000 responded to the survey.

    “A total of 143 students out of about 1,000 responded to this survey.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Premisecitation

    The accuracy of automatic detection of AI-generated text is quite good but the tools tend to produce too many false positives for texts written by nonnative speakers.

    “The accuracy of the automatic detection of AI-generated text is quite good but the tools tend to produce too many false positives for texts written by nonnative speakers (Liang et al., 2023).”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Liang et al., 2023

      Not checked in the free run.

    • Rests on an assumption: Tools exist for textual content (e.g., OpenAI AI Classifier, OriginalityAI, GPZERO) and multimedia content (e.g., Illuminarty) to detect AI-generated content.
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationreasoning

    With the proper amount of knowledge, students can utilise this knowledge to develop their critical thinking abilities.

    “With the proper amount of knowledge students can then utilise this knowledge to develop their critical thinking abilities.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Premisenone given

    Communication skills also apply to a conversation between a human and an AI.

    “Communication skills also apply to a conversation between a human and an AI.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Interpretationnone given

    A risk is that humans will rely on AI so much that they will deprive themselves of knowledge and reduce their ability to think logically and reason.

    “A risk is that humans will rely on AI so much that they will deprive themselves of knowledge and reduce their ability to think logically and reason.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

    Premisenone given

    In addition to fostering cognitive development, the education process also plays a key role in shaping interpersonal skills, emotional intelligence and ethical values.

    “In addition to fostering cognitive development the education process also plays a key role in shaping interpersonal skills, emotional intelligence and ethical values.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationreasoning

    A strong knowledge base equips individuals to detect inconsistencies or inaccuracies, thus minimising the risk of accepting false information as truth.

    “A strong knowledge base also equips individuals to detect inconsistencies or inaccuracies thus minimising the risk of accepting false information as truth.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Interpretationreasoning

    Humans are predicted to assume roles as managers and trainers of program robots, prompting a shift in demand from low-skilled white-collar employees to those possessing high-level and interdisciplinary skills required for taking care of 'program robots'.

    “Humans are predicted to assume roles as managers and trainers of these program robots prompting a shift in demand from low-skilled white-collar employees who were traditionally educated to serve clients to those possessing high-level and interdisciplinary skills required for taking care of”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Premisenone given

    In order to distinguish facts from hallucinations, an individual must possess personal knowledge.

    “In order to distinguish facts from hallucinations an individual must possess personal knowledge.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationnone given

    Creativity flourishes in an educational environment that encourages exploration and divergent thinking.

    “Furthermore, creativity which is an essential aspect of intellectual development flourishes in an educational environment that encourages exploration and divergent thinking.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationreasoning

    This challenge will increase as AI improves and makes fewer errors.

    “This challenge will increase as AI improves and makes fewer errors.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationreasoning

    The rapid development of generative AI makes it more likely that white-collar jobs are also at risk.

    “However, the rapid development of generative AI makes it more likely that white-collar jobs are also at risk.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationreasoning

    Incorporating AI literacy into higher education curricula can contribute to fostering a culture of responsible AI use.

    “Moreover, incorporating AI literacy into higher education curricula can contribute to fostering a culture of responsible AI use.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    These challenges, while demanding and time-consuming, ultimately contribute to the development of various skills and competencies.

    “These challenges while demanding and time-consuming ultimately contribute to the development of various skills and competencies.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    To achieve this, it is necessary that the educational process presents challenges that require from students significant effort and perseverance to overcome.

    “To achieve this it is necessary that the educational process presents challenges that require from students significant effort and perseverance to overcome.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    Under such circumstances it becomes exceedingly easy to make erroneous decisions which may bear significant consequences.

    “Under such circumstances it becomes exceedingly easy to make erroneous decisions which may bear significant consequences.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    The field of explainable AI, which seeks to supplement machine learning with techniques that offer insight into these correlations, is still in its infancy.

    “The field of explainable AI which seeks to supplement machine learning with techniques that offer insight into these correlations is still in its infancy.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisenone given

    The communication skills include the ability to ask the AI pertinent questions, sometimes repeatedly, sometimes phrased differently and sometimes using different keywords in order to get results that are as truthful as possible to understand the problem being described.

    “The communication skills include, in particular the ability to ask the AI pertinent questions, sometimes repeatedly, sometimes phrased differently and sometimes using different keywords in order to get results that are as truthful as possible to understand the problem being described.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    Engaging in a rigorous educational process in which individuals deal with complex ideas and problems stimulates critical thinking, problem solving and creativity.

    “Engaging in a rigorous educational process in which individuals deal with complex ideas and problems stimulates critical thinking, problem solving and creativity.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    The acquisition of facts from credible sources aids in fostering effective communication skills.

    “The acquisition of facts from credible sources aids in fostering effective communication skills.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationreasoning

    By emphasizing critical analysis of AI-generated content, academic institutions can help students develop a nuanced understanding of the strengths and limitations of AI solutions.

    “By emphasizing critical analysis of AI-generated content academic institutions can help students develop a nuanced understanding of the strengths and limitations of AI solutions.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationreasoning

    When students meet the established requirements and attain a certain level of education, they become more attractive to employers.

    “When students meet the established requirements and attain a certain level of education they become more attractive to employers.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Premisenone given

    Education plays an important role in supporting the growth and intellectual development of individuals.

    “Education plays an important role in supporting the growth and intellectual development of individuals.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    AI models, especially those based on machine learning, are founded on correlation rather than cause-and-effect relationships.

    “This is primarily because AI models and especially those based on machine learning are founded on correlation rather than cause-and-effect relationships.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    When students are presented with challenging tasks that require them to think outside the box, they can exploit their creative potential.

    “When students are presented with challenging tasks that require them to think outside the box they can exploit their creative potential.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S4 (based).

    Interpretationnone given

    The increase in productivity, global accessibility, multilingual capabilities and personalization offered by software robots is likely to reduce the overall need for human labour in the market.

    “Unfortunately the increase in productivity, global accessibility, multilingual capabilities and personalization offered by software robots is likely to reduce the overall need for human labour in the market.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Premisenone given

    Human decision-making can take a relatively long time and, despite well-organized cognitive mechanisms, humans are prone to errors.

    “However, it should be noted that the human decision-making process can take a relatively long time and despite the well-organized cognitive mechanisms humans are prone to errors.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S4 (based).

    Premisereasoning

    The most important educational challenge arising from unrestricted access to generative AI solutions such as ChatGPT is the blurring of boundaries between factual information and generated hallucinations.

    “The most important educational challenge arising from unrestricted access to generative AI solutions such as ChatGPT is the blurring of boundaries between factual information and generated hallucinations.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationreasoning

    This form of mental exercise promotes the development of cognitive abilities as well as adaptability and resilience in the face of adversity.

    “This form of mental exercise promotes the development of cognitive abilities as well as adaptability and resilience in the face of adversity.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationreasoning

    This not only contributes to personal growth but also promotes collective progress in the pursuit of knowledge and innovation.

    “This not only contributes to personal growth but also promotes collective progress in the pursuit of knowledge and innovation.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    The use of AI may be an excellent opportunity for developing creative thinking.

    “The use of AI may be an excellent opportunity for developing creative thinking.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationnone given

    One major error made by AI undetected by humans beforehand may have catastrophic consequences.

    “Still one major error made by AI undetected by humans beforehand may have catastrophic consequences.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationreasoning

    By setting specific requirements for each level of education, universities establish a standardized framework that allows for consistent measurement of knowledge and skills.

    “By setting specific requirements for each level of education universities establish a standardized framework that allows for consistent measurement of knowledge and skills.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S4 (based).

    Premisenone given

    Artificial intelligence primarily operates based on statistical patterns.

    “Conversely artificial intelligence primarily operates based on statistical patterns.”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The increasing advancement of artificial intelligence raises the possibility that sources of information may be contaminated with hallucinations.

    “Although one may peruse various sources of information for this purpose the increasing advancement of artificial intelligence raises the possibility that these sources may be contaminated with hallucinations.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Interpretationreasoning

    Only an individual equipped with knowledge can utilise artificial intelligence in a sensible manner, thus minimising the risk of confusing facts with hallucinations and drawing erroneous conclusions.

    “Only an individual equipped with knowledge can utilise artificial intelligence in a sensible manner thus minimising the risk of confusing facts with hallucinations and drawing erroneous conclusions.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Interpretationreasoning

    By adopting this mindset, individuals are more likely to persevere in their work and view setbacks as opportunities for personal growth rather than as insurmountable barriers.

    “By adopting this mindset individuals are more likely to persevere in their work and view setbacks as opportunities for personal growth rather than as insurmountable barriers.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    By engaging in cooperative learning experiences, participating in diverse learning environments and reflecting on moral and ethical dilemmas, students can develop empathy, cultural competence and a strong sense of personal responsibility.

    “By engaging in cooperative learning experiences, participating in diverse learning environments and reflecting on moral and ethical dilemmas students can develop empathy, cultural competence and a strong sense of personal responsibility.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    The potential risk for decision-makers stems from the confusion of recommendations derived from statistical analysis with those based on logical reasoning.

    “The potential risk for decision-makers stems from the confusion of recommendations derived from statistical analysis with those based on logical reasoning.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    This understanding will enable them to make informed decisions about when and how to use these tools in their academic pursuits and professional careers.

    “This understanding will enable them to make informed decisions about when and how to use these tools in their academic pursuits and professional careers.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationreasoning

    With a good understanding of a subject, individuals are better equipped to engage in meaningful conversations, exchange ideas and collaborate with others in their academic or professional fields.

    “With a good understanding of a subject individuals are better equipped to engage in meaningful conversations, exchange ideas and collaborate with others in their academic or professional fields.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationnone given

    The advent of generative AI promises to revolutionize the nature of intellectual labour, often referred to as 'white-collar' work.

    “However, the advent of generative AI promises to revolutionize the nature of intellectual labour, often referred to as”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Interpretationreasoning

    It is anticipated that the labour market will exhibit a declining demand for low-skilled production workers and an increasing demand for highly skilled engineers tasked with operating, maintaining, integrating, developing, and programming robots.

    “As this trend continues it is anticipated that the labour market will exhibit a declining demand for low-skilled production workers and conversely an increasing demand for highly skilled engineers tasked with operating, maintaining, integrating, developing, and programming robots.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    When universities establish clear and challenging requirements, students are motivated to continuously improve their knowledge and skills to meet those expectations.

    “When universities establish clear and challenging requirements students are motivated to continuously improve their knowledge and skills to meet those expectations.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable).

    Premisereasoning

    The high statistical probability of a particular phenomenon reoccurring derived from past data does not guarantee its future repetition.

    “The high statistical probability of a particular phenomenon reoccurring derived from past data does not guarantee its future repetition.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    The process of tackling difficult and complex tasks fosters the belief that abilities can be developed through effort and learning.

    “What is more the process of tackling difficult and complex tasks fosters the belief that abilities can be developed through effort and learning.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

    Premisereasoning

    An important role of universities remains to set requirements appropriate to a given level of education and verify the attainment of this level by students.

    “An important role of universities remains to set requirements appropriate to a given level of education and verify the attainment of this level by students.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Interpretationreasoning

    This may lead to the propagation of misinformation and hinder the acquisition of accurate knowledge.

    “Consequently this may lead to the propagation of misinformation and hinder the acquisition of accurate knowledge emphasising the need for critical evaluation and verification of AI-generated content in educational settings.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

    Interpretationreasoning

    All this influences the positive motivation of students to work and helps promote talented individuals.

    “All this influences the positive motivation of students to work and helps promote talented individuals.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Premisenone given

    A student enters a university when he/she is 19 years old, leaves a university when he/she is 24 years old, and has to work for 45 years-until he/she reaches the age of 69 (in the near future).

    “A student enters a university when he/she is 19 years old, leaves a university when he/she is 24 years old, and has to work for 45 years-until he/she reaches the age of 69 (in the near future).”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim. May combine two claims.

    Interpretationreasoning

    The development of mechanisms to distinguish between factual information and hallucinations becomes imperative, which currently relies on human involvement, particularly individuals who possess extensive knowledge.

    “Consequently the development of mechanisms to distinguish between factual information and hallucinations becomes imperative which currently relies on human involvement, particularly individuals who possess extensive knowledge.”
    Quote in the paperSubstantial claim
  • Classified as probably not a substantial claim.

    Interpretationreasoning

    The nuances of language, cultural references and context-specific elements may not always be accurately captured by AI systems, leading to potential misinterpretations or misunderstandings.

    “The nuances of language, cultural references and context-specific elements may not always be accurately captured by AI systems leading to potential misinterpretations or misunderstandings.”
    Quote in the paperSubstantial claim
  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Premisenone given

    Technology kills old jobs and creates new ones.

    “Technology kills old jobs and creates new ones.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationreasoning

    This creates organizational and competency gaps that continue to grow.

    “This creates organizational and competency gaps that continue to grow.”
    Quote in the paperNot a substantial claim
  • May combine two claims.

    Interpretationreasoning

    Training AI models on hallucinatory outputs may intensify the generation of such hallucinations, potentially leading to a decline in the quality of the generated texts.

    “Training AI models on hallucinatory outputs may intensify the generation of such hallucinations potentially leading to a decline in the quality of the generated texts.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationnone given

    AI can help people keep up with changes by anticipating how their work environment and required competencies will change and assisting them in the necessary development.

    “AI can help people keep up with changes by anticipating how their work environment and required competencies will change and assisting them in the necessary development.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S4 (based). May combine two claims.

    Premisenone given

    GPT models have been trained on data collected until 2021 while the corpus of available data continues to grow and evolve.

    “However, it is important to note that GPT models have been trained on data collected until 2021 while the corpus of available data continues to grow and evolve.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    No one at a university can predict what to teach students now that will be useful to them in their job 35-45 years from now.

    “No one at a university can predict what to teach students now that will be useful to them in their job 35-45 years from now, i.e., for the last ten years of their working life.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

    Interpretationnone given

    A university that sets appropriate educational requirements fosters a competitive environment among its students.

    “Moreover, a university that sets appropriate educational requirements fosters a competitive environment among its students.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

    Interpretationreasoning

    Due to possible hallucinations and uncertainty about the veracity of the knowledge offered by AI, it will be necessary to work with competent and knowledgeable people whose pool of knowledge should remain at the university.

    “Due to possible hallucinations and uncertainty about the veracity of the knowledge offered by AI however it will be necessary to work with competent and knowledgeable people whose pool of knowledge should remain at the university.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

    Interpretationreasoning

    This increases their chances of securing well-paid jobs and advancing in their careers.

    “This increases their chances of securing well-paid jobs and advancing in their careers.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable). May combine two claims.

    Interpretationreasoning

    Universities will be unable to prepare students in advance for new jobs, so they should provide students with the capabilities to quickly update and develop new skills.

    “Universities will be unable to prepare students in advance for new jobs so they should provide students with the capabilities to quickly update and develop new skills.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Premisereasoning

    Notable advancements in AI algorithms, as exemplified by the comparison between GPT-3 and GPT-4, indicate potential improvements in the quality of AI-generated content.

    “On the one hand notable advancements in AI algorithms as exemplified by the comparison between GPT-3 and GPT-4, indicate potential improvements in the quality of AI-generated content.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Premisereasoning

    While AI models aim to generate coherent and contextually relevant output, the interpretation of meaning and intent remains subjective and reliant on human understanding.

    “While AI models aim to generate coherent and contextually relevant output the interpretation of meaning and intent remains subjective and reliant on human understanding.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable), S4 (based). May combine two claims.

    Premisenone given

    The world is developing unevenly; technology and especially AI is growing exponentially, organizations are growing logarithmically, and education is growing by small leaps.

    “The world is developing unevenly.Technology and especially AI is growing exponentially, organizations are growing logarithmically, and education is growing by small leaps.”
    Quote in the paperNot a substantial claim
  • Probably not a substantial claim: may not pass S2 (checkable).

    Interpretationreasoning

    The subjective nature of interpretation introduces a level of variability among individuals as different people may perceive and understand AI-generated content in diverse ways.

    “Furthermore, the subjective nature of interpretation introduces a level of variability among individuals as different people may perceive and understand AI-generated content in diverse ways.”
    Quote in the paperNot a substantial claim
  • Classified as probably not a substantial claim.

    Premisereasoning

    In addition to concerns about the quality of AI-generated content, the subjective interpretation of such content poses another significant challenge.

    “In addition to concerns about the quality of AI-generated content the subjective interpretation of such content poses another significant challenge.”
    Quote in the paperSubstantial claim
80 of 80
  • The paper relies on this without support.

    GPT-4 can analyse as many as 32,768 tokens (about 24,000 words).

    “Reportedly, GPT-4 is able to analyse as many as 32,768 tokens (about 24,000 words).”
  • The paper relies on this without support.

    Generative AI systems are widely accessible to higher education students.

    “The availability of generative AI systems such as the GPT-4, Open Assistant, DALL-E, Midjourney, and many others has dramatically changed the landscape of information acquisition and learning.”
  • The paper relies on this without support.

    Easy access to generative AI leads to academic dishonesty.

    “Then the focus is on the threats posed by easy access to automatically generated content such as academic dishonesty and the possible erosion of critical thinking skills.”
  • The paper relies on this without support.

    Easy access to generative AI leads to the possible erosion of critical thinking skills.

    “Then the focus is on the threats posed by easy access to automatically generated content such as academic dishonesty and the possible erosion of critical thinking skills.”
  • The paper relies on this without support.

    The style of a given artist can very easily be reproduced by an AI system.

    “The style of a given artist can very easily be reproduced by an AI system (cf.Section 1.2).”
  • The paper relies on this without support.

    A person without knowledge is unable to distinguish between the right and wrong parts of the generated text.

    “A person without knowledge is unable to distinguish between the right and wrong parts of the generated text.”
  • The paper relies on this without support.

    Hallucinations are often intertwined with accurate information and conveyed in a persuasive and self-assured manner, rendering their detection difficult without careful scrutiny and rigorous fact verification.

    “Hallucinations are often intertwined with accurate information and conveyed in a persuasive and self-assured manner rendering their detection difficult without careful scrutiny and rigorous fact verification.”
  • The paper relies on this without support.

    Open-domain hallucinations are more challenging and require comprehensive research to identify.

    “Contrarily open-domain hallucinations such as those presented above are a more challenging case which requires comprehensive research to identify.”
  • The paper relies on this without support.

    AI-generated content is often indistinguishable from human-created content for a typical human reader.

    “AI-generated content is often indistinguishable from human-created content for a typical human reader.”
  • The paper relies on this without support.

    To some extent it is possible to automatically detect AI-generated content.

    “However, to some extent it is possible to automatically detect such content.”
  • The paper relies on this without support.

    Tools exist for textual content (e.g., OpenAI AI Classifier, OriginalityAI, GPZERO) and multimedia content (e.g., Illuminarty) to detect AI-generated content.

    “Tools exist for textual content (e.g., OpenAI AI Classifier, OriginalityAI, GPZERO) and multimedia content (e.g., Illuminarty).”
  • The paper relies on this without support.

    Automatic identification of AI-generated images and videos is more challenging than detecting text, but this type of multimedia content is easier to distinguish by humans.

    “Automatic identification of AI-generated images and videos is more challenging than detecting text, but at the same time this type of multimedia content is easier to distinguish by humans.”
  • The paper relies on this without support.

    Typical problems in AI-generated images include incorrect proportions of the human body, poorly represented hands and fingers and deformed background elements.

    “Typical problems in AI-generated images include incorrect proportions of the human body, poorly represented hands and fingers and deformed background elements.”
  • The paper relies on this without support.

    The survey respondents are representative of the broader student population.

    “Only 2% of students declared complete trust in the content generated by AI systems (Figure 9, left).”
  • The paper relies on this without support.

    The factual error in the GPT-3.5 biography is obvious.

    “57.5% failed to identify the obvious error in the content.”
  • The paper relies on this without support.

    The survey questions accurately measure students' trust and attentiveness.

    “First an analysis was carried out to determine to what extent students trust content generated by AI and if they can really identify obvious problems with automatically generated content.”
  • The paper relies on this without support.

    The overall sample size of the survey is sufficient to draw statistically significant conclusions.

    “The sample is too small to draw any statistically significant conclusions from this.”
17 of 17

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?

  • Excessive use of GenAI can inhibit the development of critical thinking skills and autonomy in students.

    “Another relevant aspect is cognitive dependence, where excessive use of GenAI can inhibit the development of critical thinking skills and autonomy in students (Walczak and Cellary, 2023).”

    From Ethical and regulatory challenges of Generative AI in education: a systematic review (2025) · cites this paper as Walczak and Cellary, 2023

    Section 4, “Consequences for higher education” (page unavailable), warns that relying too much on AI may reduce people’s knowledge and ability to reason, and says education must address this risk. Sections 5.1 and 5.4 connect students’ knowledge and critical-thinking development to responsible AI use. The paper does not specifically say that excessive use inhibits the development of students’ autonomy. (LLM, conf=0.84)

    Experimental — accuracy not yet measured
  • The digital divide can limit equitable access to GenAI technologies in disadvantaged contexts.

    “It is noted that the digital divide can limit equitable access to these technologies in disadvantaged contexts (Walczak and Cellary, 2023).”

    From Ethical and regulatory challenges of Generative AI in education: a systematic review (2025) · cites this paper as Walczak and Cellary, 2023

    Section 1.1.1, “Large Language Models” (page unavailable), says high training costs may exacerbate the digital divide by concentrating AI capabilities in well-funded organizations and institutions. That supports an unequal-access concern, but the paper does not specifically discuss access to GenAI technologies in disadvantaged contexts. (LLM, conf=0.83)

    Experimental — accuracy not yet measured
  • The development and operation of GenAI carry significant environmental costs, including high energy and resource consumption, and raise concerns about labor exploitation and job displacement.

    “The development and operation of GenAI carry significant environmental costs, including high energy and resource consumption, and raise concerns about labor exploitation and job displacement.”

    From MAPPING GENERATIVE AI'S ETHICAL ISSUES IN HIGHER EDUCATION: A FELT-GUIDED SYSTEMATIC REVIEW [PEMETAAN ISU ETIKA GENERATIVE AI DI PENDIDIKAN TINGGI: TINJAUAN SISTEMATIS BERPANDUAN FELT] (2025) · cites this paper as Walczak & Cellary, 2023

    Section 4, “Consequences for higher education” (page unavailable), says generative AI is likely to reduce demand for human labour and put white-collar jobs at risk. The Introduction mentions increasing computational power, but does not establish significant environmental costs or high energy and resource consumption; the searched passages also do not address labour exploitation. (LLM, conf=0.88)

    Experimental — accuracy not yet measured
  • Walczak and Cellary's (2023) study found that 50% of participants used generative AI for writing-based tasks.

    “This supports Walczak and Cellary's (2023) study that found 50% of participants using generative AI for writing-based tasks.”

    From A Survey on Student Use of Generative AI Chatbots for Academic Research (2024) · cites this paper as Walczak and Cellary's (2023)

    The survey results (section 3.1, page unavailable) report students’ use of AI, and the passage headed “out of curiosity” lists generating an essay or abstract among its uses. It does not report that 50% of participants used AI for writing-based tasks; it describes essay generation as a less frequently selected use. (LLM, conf=0.86)

    Experimental — accuracy not yet measured
  • There is a need for support in new digital literacy.

    “Walczak and Cellary [60] also highlighted the need for support in this new digital literacy.”

Cites

Flagged first, then by citations

Research works printed but not linked: 2

  1. Linguistics: Human Language Technologies, 1, 4171–4186. Minneapolis, Minnesota. Association for Computational Linguistics.
  2. Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A., & Bengio, Y. (2014). Generative adversarial nets. Advances in Neural Information Processing Systems, 27.

Cited by

Sorted by citations

Authors
Walczak, K., & Cellary, W.
Affiliations
  • Poznań University of Economics and Business (PL)
  • WSB Merito University in Poznan (PL)
Venue
Economics and Business Review
Published
2023-07-20
Volume
9
Issue
2
ISSN
  • 1643-5877 (ISSN-L)
  • 2450-0097
  • 2392-1641
OpenAlex
W4385070262
Cited by
196
References
31
Citations by year
2026: 52 · 2025: 86 · 2024: 57 · 2023: 1
FWCI
46.00
Type
journal article
Language
en
Open access
Locations
  • Economics and Business Review/The Poznań University of Economics Review (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Physical Sciences › Computer Science › Information Systems › Artificial Intelligence in Education
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Physical Sciences › Computer Science › Information Systems › Educational Innovations and Challenges
Keywords
generative AI, higher education, AI in teaching, GPT-4, digital literacy, AI ethics, curriculum change, student perceptions of AI, role of educators, AI policy in 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.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-09.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-09.

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