Abstract 1. Introduction
Generative artificial intelligence (AI) has increasingly become part of adolescents' everyday experiences and is being used for learning, content creation, problem-solving, entertainment, and emotional interaction. Despite its potential to improve access to information and support cognitive processes, its use may also be associated with reduced cognitive effort, dependency, misinformation, diminished creativity, and ambiguity regarding the boundary between human and machine interaction. The present study aimed to understand how Iranian adolescents make meaning of their experiences of interacting with generative AI and to identify strategies for balanced and bounded use of this technology.
2. Methodology
This qualitative study employed a reflexive thematic analysis approach. Participants were 14 adolescents aged 14–18 years from Tehran Province who were selected through purposive sampling. Eligibility criteria included being an adolescent, having direct experience using at least one generative AI tool, and being willing to describe that experience. Participants were recruited through schools in coordination with relevant school personnel. Data were collected through qualitative interviews and analyzed through an iterative process involving movement between the data, codes, and themes. To enhance the trustworthiness of the findings, the criteria of credibility, dependability, confirmability, and transferability were considered throughout the research process.
3. Findings
The analysis resulted in six main themes: (1) efficiency and organization of everyday activities, (2) blurring of instrumental and human positions, (3) productivity at the cost of cognitive effort, (4) conditional trust and preservation of human authority, (5) keeping AI in its appropriate place, and (6) a dual-sided future. The findings indicated that adolescents used generative AI for learning, summarization, planning, coding, content creation, and emotional support. They valued its speed, constant accessibility, and nonjudgmental interaction. At the same time, they expressed concerns about replacing personal effort with AI, reduced creativity, dependency, misinformation, and threats to privacy. Their trust in AI was conditional and varied according to the nature and potential consequences of the task. In areas such as physical and mental health, adolescents maintained the importance of human experts. They also established personal boundaries regarding the amount and type of AI use to preserve their autonomy, creativity, and privacy. However, fatigue, time pressure, and academic demands sometimes led them to cross these boundaries.
4. Discussion and Conclusion
The findings were integrated into the central theme of "effective, bounded interactions." This theme indicates that adolescents' experiences with generative AI cannot be reduced to a simple acceptance–rejection dichotomy. Rather, the consequences of AI use depend largely on its position within the adolescent's cognitive and relational processes. When AI is used after an initial effort, it can function as a cognitive scaffold and provide constructive feedback; when used before personal effort, it may replace learning and weaken cognitive agency. Although the natural language of AI can create a sense of social presence and being understood, adolescents still distinguish between simulated empathy and genuine human emotional experience. Thus, productive use of AI does not necessarily require avoiding the technology; instead, it involves maintaining AI in the position of a supporting tool rather than a substitute for human thinking and judgment. AI literacy education in schools, particularly through the strategy of "effort–question–verification," along with teaching information verification, privacy protection, and appropriate referral to human experts, may contribute to more balanced use among adolescents. The study was limited by its small purposive sample, reliance on self-reported data, and the absence of direct observation of adolescents' actual interactions with AI systems. Future research could use more diverse samples, longitudinal designs, and the perspectives of adolescents, parents, and teachers to examine how adolescents' patterns of AI use and boundary-setting change over time.
