ChatGPT's Essay on Kohlberg's Theory: AI's Use in Academic Writing
The following human-created article assesses an AI-generated essay, and this page, therefore, does not contain AI-generated content. The ChatGPT-generated article is linked here.
Aaqib F. Azeez, January 2026
Abstract: The usage of artificial intelligence has raised significant questions regarding its accuracy and reliability. This paper assesses an essay created by ChatGPT (Model 5.2) on Lawrence Kohlberg's Cognitive-Developmental Theory regarding moral development and modern viewpoints on the theory. The AI-generated essay demonstrated several strengths, including accurately describing Kohlberg's stages and using authentic and academically-appropriate sources. Despite the positives, the essay also had significant drawbacks, including a lack of citations for strong claims. The findings in the paper point to several positive contributions AI can have in academic and professional tasks, but this should be treated with caution and preferably overseen by competent professionals.
Keywords: artificial intelligence, ChatGPT, academia, workforce, Lawrence Kohlberg's stages of moral development, citation accuracy
 |
Educational level: this is a tertiary (university) resource. |
 |
Subject classification: this is a psychology resource. |
 |
Type classification: this is a paper resource. |
 |
Completion status: this resource has reached a high level of completion. |
Introduction
[edit | edit source]The usage of AI has caused a lot of chaos in the academic world. One of the issues with the usage of AI is the accuracy of its work. Although AI is convenient for brainstorming ideas or providing a framework to work and improve on, using AI for a finished product is counterintuitive, since AI can hallucinate details to please the user. The purpose of this paper is to assess the accuracy of AI by evaluating an essay made by the ChatGPT model 5.2 on Lawrence Kohlberg’s Cognitive-Developmental Theory.
Critical analysis
[edit | edit source]The essay created by ChatGPT hosts a variety of positives. Firstly, the in-text citations correctly correspond to the references listed at the bottom of the page. For example, “Killen & Dahl, 2021” correctly corresponds to the reference “Killen, M., & Dahl, A. (2021)” listed under “References”. Secondly, all the publications ChatGPT used were published later than 2018, according to the instructions given to the AI model. Thirdly, the in-text citations were accurate to the sources. For example, the paper cites Mammen & Paulus (2023) to support the assertion that moral reasoning should be examined under natural conversations with other people, and not just a structured interview[1]. Mammen & Paulus (2023) support the AI’s statements since the original paper notes that interviews can only study the final product of decision-making, therefore missing perceptions on the process of decision-making[1]. Mammen & Paulus (2023) upheld the importance of moral development and its origin “rooted in human communication.”[1] The essay provided by the AI is also accurate, correctly identifying each stage of Kohlberg’s Cognitive-Developmental Theory. The essay mentions that the preconventional level is characterized by moral judgment that focuses on “punishment avoidance or personal benefit”, the conventional level is a witness to the beginning of “social approval, rules, and [the maintenance of] social order” influencing morality, and the postconventional stage being the final level where morality ascends to the “broader principles such as justice, rights, and fairness that may sometimes conflict with authority”. Santrock (2025) describes Kohlberg’s first level in a similar way, painting the preconventional stage as being “strongly influenced by external punishment and reward”, describing the conventional reasoning stage as the site of “understand[ing] the importance of following the laws of society”, and reports the postconventional reasoning stage as the “highest” level, where norms are being pitted against “moral concerns such as liberty, justice, and equality, with the idea that morality can improve the laws” (p. 228)[2]. Lastly, the AI-generated essay stays on topic and doesn’t deviate from the main subject and provides the readers with up-to-date information on the subject. The essay discusses the basics of Kohlberg’s cognitive-developmental theory (including the 3 stages) and includes modern developments of Kohlberg’s theories, including the integration of individual differences, culture, social contexts, and emotions.
Although the essay is largely accurate and useful, it is marred by a few shortcomings. The AI-generated essay makes the claim that a “common critique” of Kohlberg’s theory is that Kohlberg’s description of “advanced reasoning” is not always practical in “real-world moral decisions, especially under stress or social pressure.” Since the essay claims that this is a “common critique”, a citation to back this up should be provided. Another instance of this is where ChatGPT claims that the stage theories have received criticism for “assumptions about universality and the primacy of justice-based reasoning.” The same issue is repeated here, where there are no citations or sources provided to justify this claim. The lack of sources for broad claims that the AI has made does not decimate the credibility of the essay but hampers the ability to verify and fact-check certain statements that hold modern weight.
Usage of AI in the real world
[edit | edit source]The addition of AI in professional settings can compensate for deficiencies in the speed at which humans collect information. Notable positive traits of AI include its speedy collection and examination of data, the ability to improve with minimal human aid after training, and the ability to carry out crucial decision-making processes based on analytical reasoning[3]. A 2021 Deloitte and MedTech Europe report projected that AI could potentially save 380,000-403,000 lives per year in European healthcare[4]. Even from the AI-generated essay, AI can not only provide a well-written description of an entity but also include accurate citations that correspond with the text. The combination of AI’s analytical pace and the human’s heightened judgement and integration of “abstract thinking” and “intuitive approach” could bolster each other’s performance in certain tasks, as observed in a 2016 study reported by Jarrahi (2018) that saw an 85% reduction in error in cancer detection when AI and pathologists both collaborated on the task[3]. With AI’s ability to process and analyze information efficiently and quickly, it can prove to be a handy tool that can bolster human production.
Despite AI’s efficient processing of information, AI has been notorious for hallucinating information, as was observed in a 2023 incident where a group of lawyers from Levidow, Levidow & Oberman, P.C. were caught publishing fabricated court decisions marred with “fake quotes and citations created by the artificial intelligence tool ChatGPT”[5]. Other issues include a lack of security for confidential data, a lack of accountability for when harm may be produced by AI, and the increased likelihood of “engaging in unethical behavior” when using AI[6]. Although the AI-generated essay was largely positive and the citations were accurate, there were certain claims that were not backed up by a source. As mentioned earlier, the essay claimed that Kohlberg’s theories were criticized for their lack of practicality in real-world scenarios yet provided no citation to back this claim. It is worthwhile that when engaging with AI, it can be used for “assistance”, but oversight of such work should be done to “ensure [its] accuracy”[5].
Conclusion
[edit | edit source]The AI-generated essay was, for the most part, exceptional. The essay stayed on topic, provided accurate developments to theory with respect to the modern era, and contained accurate citations and references, but failed to back up a couple of strong claims. AI has shown itself to be quick in collecting and processing information, but supervision over its work should be done to ensure accuracy. AI should be introduced into professional settings as it clearly is a useful tool, but the notion of AI technology ‘replacing’ humans in the workforce is unfounded and would lead to more harm than commendable.
See also
[edit | edit source]- Azeez, A. (2026). ChatGPT's Essay on Kohlberg's Theory: AI's Use in Academic Writing. Zenodo. https://doi.org/10.5281/zenodo.18135581- APA-compliant paper
- Academia link
- Best Practices for Using AI in Academic Research - Institute for Humane Studies
References
[edit | edit source]- ↑ 1.0 1.1 1.2 Mammen, Maria; Paulus, Markus (2023-04-01). "The communicative nature of moral development: A theoretical framework on the emergence of moral reasoning in social interactions". Cognitive Development 66: 101336. doi:10.1016/j.cogdev.2023.101336. ISSN 0885-2014. https://www.sciencedirect.com/science/article/pii/S0885201423000412.
- ↑ Santrock, J. W. (2025). Adolescence (19th ed.). McGraw Hill Higher Education
- ↑ 3.0 3.1 Jarrahi, Mohammad Hossein (2018-07-01). "Artificial intelligence and the future of work: Human-AI symbiosis in organizational decision making". Business Horizons 61 (4): 577–586. doi:10.1016/j.bushor.2018.03.007. ISSN 0007-6813. https://www.sciencedirect.com/science/article/pii/S0007681318300387.
- ↑ Dantas, C., Mackiewicz, K., Tageo, V., Jacucci, G., Guardado, D., Ortet, S., Varlamis, I., Maniadakis, M., De Lera, E., Quintas, J., Kocsis, O., & Vassiliou, C. (2021). Benefits and hurdles of AI in the workplace – what comes next? International Journal of Artificial Intelligence and Expert Systems, 10, 9-17. https://www.researchgate.net/publication/351993615_Benefits_and_Hurdles_of_AI_In_The_Workplace_-What_Comes_Next
- ↑ 5.0 5.1 "Lawyers submitted bogus case law created by ChatGPT. A judge fined them $5,000". AP News. 2023-06-22. Retrieved 2026-01-02.
- ↑ Trincado-Munoz, Francisco J.; Cordasco, Carlo; Vorley, Tim (2025-04-26). "The dark side of AI in professional services". The Service Industries Journal 45 (5-6): 455–474. doi:10.1080/02642069.2024.2336208. ISSN 0264-2069. https://doi.org/10.1080/02642069.2024.2336208.