Abstract:

Generative artificial intelligence (GAI) involves creating models capable of synthesizing novel data resembling the training data, through techniques like generative modeling that analyze examples to discern patterns for realistic replication [1]. GAI models enable automated generation across applications including media, language, and music. Recent GAI advancements exhibit potential for transforming conventional systems, notably automatic text generation via chatbots leveraging large language models (LLMs) for natural conversation [2]. The LLM as ChatGPT from OpenAI represents optimized evolution of Generative Pre-trained Transformer (GPT) models for dialogue through conversational fine-tuning [3], [4]. Early findings indicate ChatGPT and similar LLMs may automate customer service, content creation, and other text-generation applications.

Recent studies reveal ChatGPT brings both opportunities and challenges to higher education. Firat [5] notes it can assist self-directed learning pending responsible use guidelines. Sullivan et al. [6] find media coverage concentrates on academic integrity risks over learning potential. Meanwhile, Rasul et al. [7] propose balancing adaptive learning advantages against inaccuracies and biases. Additionally, ChatGPT supports multifaceted teaching and learning via personalized tutoring, class content generation, and facilitating collaboration. It further aids research through literature reviews, data analysis, and drafting [8]. A 2022 UNESCO report highlights ChatGPT’s latent capability to enhance learning [9].

This research focuses on exploring the applications of ChatGPT in higher education settings as outlined in the UNESCO report, also considering how the students considered the way to approach AI in their professional and personal future. The study utilized a quantitative, cross-sectional survey design to investigate university students’ perspectives regarding the use of ChatGPT in higher education. Drawing from the guidelines proposed in the [9] on ChatGPT applications in academia, a structured self-report questionnaire consisting of Likert scale items was developed and administered to 140 undergraduate students from different management universities. The recruitment of student participants from diverse academic disciplines allowed for comparative analysis of any variances in viewpoints across fields of study. The sample consisted of 42% male and 58% female participants. The majority (61%) were in their first year of studies.

The preliminary results reveal interesting insights into students' perspectives on AI. On average, students tended to strongly agree that AI will have a significant presence in both their personal and professional lives in the near future (3.73(0.78)). In the academic context specifically, participants expressed a preference for utilizing ChatGPT as a search engine (3.11(0.98)) to explore and interpret data (2.93(0.83)), rather than as personal tutor or a colleague for studies. The current findings provide an initial snapshot of students' views, highlighting their openness to AI alongside the need for guidance on its conscientious use in learning and research.

Keywords:

ChatGPT, Generative AI, Higher Education.