Artificial intelligence (AI) has the potential to transform education for students with Autism Spectrum Disorders (ASD) by providing tailored support that nurtures their potential and well-being (Lampos et al., 2021). For example, AI can be effectively utilized to ensure that content is delivered in a way that resonates with the needs of students with ASD, to support communication and to enable understanding of social cues in a controlled, non-threatening setting (Adako et al., 2024). ChatGPT, as an advanced AI conversational model, can play a transformative role in educational settings. By supporting personalized learning, fostering creativity, and promoting engagement, it can cater to diverse student needs, including students with ASD (Mehta et al., 2023).
The benefits of ChatGPT include the provision of personalized learning support, since it can adapt responses to match students’ levels of understanding, enhance student engagement trough role playing, simulations and open-ended discussions, as well as support writing and idea generation (Rakap & Balikci, 2024). For students with ASD in particular, ChatGPT may offer unlimited practice opportunities in areas where students need reinforcement and repetition, without judgment (Rakap, 2024). The education systems’ support structures for students with ASD, and technology adoption levels though vary significantly across Europe (Miller, 2024), despite the existence of EU Policy Frameworks, such as the Digital Education Action Plan (2021–2027) and the European Disability Strategy 2021–2030, as well as the UNESCO Framework on AI in Education.
Based on the above EU and international frameworks, we aimed to explore how ChatGPT can enhance the inclusion and well-being of students with ASD, aligning with EU commitments to accessibility and inclusivity. In this way, we aimed to contribute to the efforts for ethical and effective use of AI in European classrooms, informing how AI integration in diverse educational contexts, may foster inclusive practices and understanding of the needs of students with ASD. We assumed that ChatGPT could assist teachers in understanding the needs of students with ASD by analyzing the type and nature of questions students ask. We propose that this analysis can offer valuable insights into the students’ cognitive preferences, challenges, and areas of interest.
In particular, students with ASD often display a preference for precise, detailed information (Sideraki et al., 2021). Hence, frequent requests for step-by-step instructions may suggest a need for structured teaching methods. In addition, questions that emphasize rules, logic, or predictable patterns can highlight a preference for structured and logical subjects (Mehta et al., 2023). Repeated inquiries on a particular topic may indicate areas where the student struggles with comprehension or requires reinforcement and point to challenges in conceptual thinking, suggesting a need for concrete examples or visual aids. If students frequently ask about specific areas, these could represent special interests, which may be leveraged by teachers to engage students more effectively in other areas of learning (Li et al., 2024). Questions about understanding emotions, social interactions, or navigating relationships may highlight areas where the student seeks support in developing social skills. Inquiries about managing stress, focus, or sensory overload may suggest specific challenges that require accommodations, such as quiet spaces or mindfulness techniques (Habibi et al., 2025).
Based on the above, our key research questions were the following:
1. What are the recurring themes in the questions students with ASD ask ChatGPT?
2. Can students’ interactions with ChatGPT provide insights into the emotional well-being and unmet needs of students with ASD?
3. What is the potential of ChatGPT to support the goals of European Union (EU) policies on inclusive education and digital transformation? 

This was a qualitative study. The participants were six students of the Special Unit from a high-school in Cyprus, one teacher and three school assistants, who were purposefully selected to participate in the study. The students were all boys with ASD, with or without comorbidities, aged between 14-17 years old. The teachers and the three school assistants were women. To safeguard anonymity of the participants, demographic details are not presented in this study. Before the implementation of the study, the parents’ informed consent was taken. In addition, the head-teacher of the school gave her permission to implement the intervention in the Special Unit, since the head-teacher is the main person that is responsible for the content and the implementation of the curriculum of the Special Unit. Moreover, the teacher and the school assistants were trained by the researcher, so as to guide students through their inquiries. The intervention included an introduction to the ChatGPT tool and the way it could be used by students. There were enough PCs for all the students in the classroom, who were encouraged to use ChatGPT to ask questions to which they needed answers. The students were also encouraged to continue asking questions, if they needed further clarifications. The school assistants, the teacher and the researcher supervised the above procedure. Since the researcher was a member of the school staff, the students were not annoyed by her presence in the classroom. The intervention lasted for approximately half an hour. The teacher continued to use the tool for the purposes of her lesson some other days, too. The main tools to gather data were the participative observation and the interview with the students, the teacher and the school assistants. Logs captured the questions asked, responses given, and patterns of interaction. Interview questions were about the students’ comfort with ChatGPT, how they felt when using it, and its perceived usefulness. Besides the content of the students’ queries, observations focused also on changes in students’ behavior, communication, or emotional states. Data were analyzed with thematic analysis. The study was compliant with research ethics: anonymity and confidentiality were kept, while both parents and students gave their informed consent to participate. Hence, this research design both provided a framework to explore the role of ChatGPT in understanding the needs of students with ASD and ensured a balance between scientific rigor and ethical responsibility.

The first theme was students’ interactions with ChatGPT. It was found that there were recurring themes in their inquiries, preference for detailed and structured information, repeated focus on specific topics or areas of interest (e.g. cars), and questions related to social skills, emotional understanding, or coping strategies. Thus, it may be assumed that the above patterns in the questions reflected the students’ cognitive preferences, learning challenges, and emotional needs. Secondly, regarding emotional well-being and behavioral insights, ChatGPT served as a safe and non-judgmental medium for students to express emotional concerns or seek help. Observations highlighted changes in students’ emotional behavior, such as increased confidence or reduced anxiety during and after their interactions with ChatGPT. In addition, there were minor communication improvements, evident in students articulating their thoughts and asking questions in a more organized manner. The third theme referred to the insights for teachers and school assistants. It was found that the analysis of question logs and feedback from students helped them better understand the needs and preferences of students with ASD. As a result, targeted interventions were designed and teaching strategies were tailored, so as to incorporate instructions, aids and guidance, based on the students’ queries. Since this study found high levels of comfort and acceptance of ChatGPT by students with ASD, due to the tool’s predictability, non-threatening nature, and instant feedback, it highlights the usability and usefulness of ChatGPT in classroom settings and indicates that it may be a helpful resource for supporting social-emotional learning and coping strategies for students with ASD. It seems then that this study supports the EU goals by fostering accessibility, inclusivity, and digital transformation through the controlled use of AI. In addition, this study may have cross-cultural applicability, by providing a model for integrating AI into special education in other European contexts.

Adako, O. P., Adeusi, O. C., & Alaba, P. A. (2024). Revolutionizing Autism Education: Harnessing AI for Tailored Skill Development in Social, Emotional, and Independent Learning Domains. Journal of Computational and Cognitive Engineering, 3(4), 348-359. Habibi, F., Sedaghatshoar, S., Attar, T., Shokoohi, M., Kiani, A., & Malek, A. N. (2025). Revolutionizing education and therapy for students with autism spectrum disorder: a scoping review of AI-driven tools, technologies, and ethical implications. AI and Ethics, 1-16. Lampos, V., Mintz, J., & Qu, X. (2021). An artificial intelligence approach for selecting effective teacher communication strategies in autism education. NPJ Science of Learning, 6(1), 25. Li, G., Zarei, M. A., Alibakhshi, G., & Labbafi, A. (2024). Teachers and educators’ experiences and perceptions of artificial-powered interventions for autism groups. BMC psychology, 12(1), 199. Mehta, P., Chillarge, G. R., Sapkal, S. D., Shinde, G. R., & Kshirsagar, P. S. (2023). Inclusion of Children With Special Needs in the Educational System, Artificial Intelligence (AI). In AI-Assisted Special Education for Students With Exceptional Needs (pp. 156-185). IGI Global. Miller, J. A. (2024). Strengthening the rights of persons with disabilities in Europe to access goods and services. European journal of legal studies, 15, 171-211. Rakap, S. (2024). Chatting with GPT: Enhancing individualized education program goal development for novice special education teachers. Journal of Special Education Technology, 39(3), 339-348. Rakap, S., & Balikci, S. (2024). Enhancing IEP Goal Development for Preschoolers with Autism: A Preliminary Study on ChatGPT Integration. Journal of Autism and Developmental Disorders, 1-6. Sideraki, A., & Drigas, A. (2021). Artificial Intelligence (AI) in Autism. Technium Science - Journal of Social Sciences, 26, 262.