ChatGPT and cognitive musical knowledge in Hindustani music

 Introduction

Generative artificial intelligence has rapidly entered higher education and is changing the way students, teachers, researchers, and professionals search for information and complete academic tasks. ChatGPT is now commonly used for explanation, content development, research support, writing, summarisation, and problem solving. Research on artificial intelligence in education suggests that these tools can provide personalised support and improve learning efficiency, but their educational value depends strongly on how they are integrated into teaching and learning. Recent research increasingly argues that AI should support human learning rather than replace human judgement and independent thinking. This issue becomes particularly important in Hindustani classical music. Musical knowledge in this tradition cannot be reduced to written definitions or lists of notes. Understanding a raga involves Swar, intonation, melodic movement, phrase structure, ornamentation, Aalap, Taan , Nyas, time association, repertoire, stylistic practice, and aesthetic interpretation. The practical knowledge developed through listening, Riyaz, performance, memory, and interaction with a teacher is therefore an important part of musical learning.

The present study examines how musicians, learners, teachers, and other music-related users perceive ChatGPT in relation to musical knowledge and cognitive musical learning. The study is based on an exploratory questionnaire involving 179 respondents. It focuses particularly on the use of ChatGPT for theory, research, raga information, practical musical analysis, creativity, and academic work. The purpose is not to argue that ChatGPT is harmful or that it reduces human intelligence. Instead, the study investigates an important educational question: whether frequent AI use changes the amount of independent thinking, verification, musical judgement, and knowledge construction the learner undertakes.

The study also considers recent research on AI literacy, prompt engineering, hallucination, feedback, and human-AI interaction to understand the survey findings within a wider educational context.

Survey analysis

The survey included 179 respondents. Because demographic information, sampling procedures, geographical distribution, complete questionnaire details, and validation statistics were not available, treat the findings as exploratory descriptive evidence rather than representative evidence of the entire music community. Percentages were calculated from the reported response frequencies, and no inferential or causal claims are made.  The first important finding concerns the purposes for which respondents use ChatGPT. A total of 105 respondents (58.7%) reported using ChatGPT for theory knowledge and content searching. Seventy respondents, or 39.1%, used it for research work, while 4 respondents, or 2.2%, reported using it mainly for lyrics and music content. 

The findings therefore show that ChatGPT is primarily being used as an information and academic support tool. Its role is less about replacing musical practice and more about speeding up information and content development. However, the respondents were less confident about the quality of AI-generated information. Only 45 respondents, or 25.1%, considered ChatGPT results accurate. Ninety respondents, or 50.3%, considered the results average, while 44 respondents, or 24.6%, considered them unsatisfactory or surface level. This finding matters because it shows a gap between access and trust. Users may find ChatGPT useful even when they do not completely trust its answers.

The concern becomes stronger when raga information is considered. One hundred and thirty respondents, or 72.6%, stated that ChatGPT provides information about ragas but that its accuracy requires verification. Only 49 respondents, or 27.4 %, considered the information complete and reported relying on it. At the same time, 160 respondents, or 89.4%, reported using ChatGPT for raga analysis and practical musical information. Only 19 respondents did not use it for these purposes.  This is the survey’s central finding. ChatGPT is widely used for raga analysis, but most users do not consider its output sufficiently reliable without verification. Findings on deeper musical analysis are similar. Only 67 respondents, or 37.4%, considered ChatGPT analysis satisfactory, whereas 112 respondents, or 62.6%, considered it unsatisfactory.  The survey also shows that conventional academic resources remain important. One hundred and forty-five respondents, or 81 %, reported continued dependence on question banks, classroom notes, presentations, email, reference notes, and other academic resources. The primary motivation for AI use was efficiency. One hundred and nineteen respondents, or 66.5 %, identified reference work, content creation, and time saving as major reasons for using ChatGPT. Another 60 respondents, or 33.5 %, identified formatting and time saving. 

Finally, perceptions about creativity were divided. Sixty-five respondents, or 36.3 %, believed that ChatGPT enhances creativity and knowledge. Thirty-three respondents, or 18.4%, believed it harms creativity or cognitive activity. The largest group, 81 respondents (45.3%), viewed ChatGPT mainly as a time-saving tool and remained uncertain about its cognitive effects. 

Results and discussion

The survey demonstrates that the relationship between ChatGPT and musical learning is not simply positive or negative. The most visible benefit is efficiency. ChatGPT can help users organise information, prepare academic material, generate ideas, improve language, and identify possible research directions. Similar benefits have been reported in other areas of higher education, where structured AI use improved language, organisation, and content in academic writing while also increasing students' confidence in using effective prompts.  Wang, Yin, and Cao similarly demonstrated that ChatGPT can be effective when integrated into active, inquiry-based, and adaptive learning. Their quasi-experimental study showed improvements in student learning outcomes and motivation when AI was used within a structured pedagogical framework. This finding is directly relevant to music education. AI may be useful when students question, compare, verify, and reflect rather than simply accept generated answers. Reliability is equally important. Recent research shows that large language models can vary in reliability depending on the task, prompt, and evaluation criteria. Liu, Ye, and Yan found substantial variation between large language models in educational assessment, including differences in consistency and alignment with expert human judgements. 

This supports the present survey finding that users may use ChatGPT without regarding it as a final authority. Verification is particularly important for Hindustani music. A raga description may appear convincing while still missing stylistic or contextual distinctions. AI can provide a theoretical description, but the learner must determine whether it aligns with recognised musical sources and actual performance practice. This is where subject knowledge becomes essential. Research on prompt engineering also supports this argument. Pangestu et al. reviewed 76 studies and identified prompt engineering as an important component of AI literacy in higher education. Their review found that students need to learn how to formulate contextual prompts, request explanations, and verify AI-generated information. The review further notes that weak prompts can lead to inaccurate or misleading responses, particularly when users have limited AI literacy. For music education, this suggests the need for what may be described as musical AI literacy. A student should understand not only how to ask ChatGPT a question, but also how to evaluate the answer musically. The student should be able to distinguish between a plausible statement and an authoritative musical statement. The risk of hallucination makes this especially important. A systematic review of 64 empirical studies by Adejumo et al. found that inaccurate or misleading Gen-AI outputs can increase cognitive difficulties and encourage overreliance. At the same time, when AI is used within structured learning environments, verification and self-monitoring can strengthen problem-solving and learning outcomes. This balanced finding is closely reflected in the present survey. Respondents use ChatGPT extensively, but many also know they need to check its answers.

Another important issue is whether students can independently judge AI output. Nazaretsky, Gabbay, and Käser compared AI- and human-generated feedback and found that students' evaluations were strongly influenced by the source’s perceived credibility. Their study also found that students could struggle to judge AI-generated feedback objectively. This is relevant to music because a learner may accept an answer because it sounds fluent or authoritative rather than because it has been musically verified. The educational implication is therefore clear. AI should be integrated through a human-in-the-loop model. Qin's institutional case study of Lingnan University similarly emphasises critical judgement, ethical reasoning, cultural values, and human competencies alongside digital fluency. The same principle applies to Hindustani music. AI can support research and learning, while teachers, performers, recordings, traditional texts, and musical practice remain important sources of validation.

Conclusion

The present survey provides evidence of extensive ChatGPT use among 179 music-related respondents. The strongest finding is the gap between use and trust. Although 89.4% use ChatGPT for raga analysis and practical musical information, 72.6% believe its Raga information requires verification. Similarly, 62.6% consider its musical analysis unsatisfactory.  The findings do not demonstrate that ChatGPT reduces human intelligence or musical creativity. They demonstrate that AI use creates a new educational responsibility. The learner must remain actively involved in questioning, listening, comparing, verifying, practising, and interpreting.

For Hindustani music, this responsibility matters because musical knowledge is embodied and contextual. A machine may generate information about a Raga, but musical understanding develops through sustained Riyaz, listening, performance, memory, aesthetic judgement, and guidance. Therefore, treat ChatGPT as a cognitive aid rather than a cognitive substitute. Its greatest educational value lies in helping learners access and organise information while encouraging stronger independent judgement. The future of AI-supported Hindustani music education should consequently focus not on rejecting AI, but on developing informed musicians who know how to use AI, question AI, verify AI, and ultimately make musical decisions independently.

References

·      Adejumo, A. A., Oyelere, S. S., Sanusi, I. T., & Suhonen, J. (2026). A systematic review of the impact of GenAI on learning performance, AI hallucinations, and problem solving in computer science education. Computers and Education: Artificial Intelligence, 10, 100570. https://doi.org/10.1016/j.caeai.2026.100570. 

·      Nazaretsky, T., Gabbay, H., & Käser, T. (2026). Can students judge like experts? A large-scale study on the pedagogical quality of AI and human personalised formative feedback. Computers and Education: Artificial Intelligence, 10, 100533. 

·      Pangestu, F. N. N., Hamidah, I., Widaningsih, L., Abdullah, A. G., & Saputra, N. A. (2026). A scoping literature review of prompt engineering for bridging students' AI literacy in higher education. Computers and Education: Artificial Intelligence, 10, 100581. https://doi.org/10.1016/j.caeai.2026.100581. 

·      Papanastasiou, E. C., Pittas, E., Rodosthenous-Balafa, M., & Lampropoulos, G. (2026). AI-assisted writing and the impact of ChatGPT on Greek-speaking students with and without learning disabilities. Outcomes from a repeated measures design. Computers and Education: Artificial Intelligence, 10, 100603. 

·      Qin, S. J. (2026). AI for education: The digital transformation of a liberal arts institution. Implementation at Lingnan University. Computers and Education: Artificial Intelligence, 10, 100592. 

·      Swaraj, A., & Chavan, P. (2024). A study on the influence of ChatGPT usage on human cognitive thinking and creativity. Library Progress International, 44(3), 4376 to 4385. The study highlights the need to balance AI-supported productivity with human creativity, critical thinking, and independent judgement. 

·      Wang, R., Yin, Y., & Cao, Y. (2026). An LLM-based pedagogical framework for active, inquiry-based and adaptive learning in L2 writing. Computers and Education: Artificial Intelligence, 10, 100535. https://doi.org/10.1016/j.caeai.2025.100535. 

·      Feng, S., & Carolus, A. (2026). Artificial intelligence literacy at school: A systematic review with a focus on psychological foundations. Computers and Education: Artificial Intelligence, 10, 100551. 

·      Long, D., & Magerko, B. (2020). What is AI literacy? Competencies and design considerations. Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems, 1 to 16. https://doi.org/10.1145/3313831.3376727. 

Share: