Capabilities and Limitations of Artificial Intelligence in Persian Audiobook Production: A Qualitative Study Based on Expert Perspectives and Sample Analysis

Capabilities and Limitations of Artificial Intelligence in Persian Audiobook Production: A Qualitative Study Based on Expert Perspectives and Sample Analysis

Fatemeh Malmir1 Mohammad Akhgari2 Mohammad Ebrahim Sadeghi Alavijeh3

1) Master s student in Radio (specializing in editorial and production) at the IRIB University, Tehran.
2) Associate Professor, Department of Radio and Television Production, IRIB University, Tehran, Iran
3) Assistant Professor, Department of Media Technical and Engineering, IRIB University, Tehran, Iran

IOI Code : AAHD-ABKKZK
Publication : National Conference on Information Technology, Nanotechnology, Artificial Intelligence and Technological Futures Studies(itnaf.ir)
Abstract :
    Abstract Advances in artificial intelligence technologies and text-to-speech systems have created new opportunities for the automatic and semi-automatic production of audiobooks. However, the application of these technologies to Persian presents technical, linguistic, and narrative challenges. This study aims to identify and analyze the capabilities and limitations of artificial intelligence in the production of Persian audiobooks. Adopting a qualitative approach, the study drew on semi-structured interviews with 27 experts in the fields of audiobooks, speech and narration, artificial intelligence, and speech processing, alongside the analysis of five audiobook samples produced using artificial intelligence. The data were analyzed through directed qualitative content analysis with a deductive orientation. The findings indicated that increased production speed, scalability, facilitation of text preparation, and reduction of some costs and repetitive tasks are among the main capabilities of artificial intelligence in this field. In contrast, limitations in emotional expression, contextual interpretation, control of tone and prosody, characterization, and the pronunciation of certain Persian linguistic structures were among the major challenges identified. The results also showed that, in the samples examined, artificial intelligence performed more effectively in simple and information-oriented texts, while the need for human supervision and evaluation increased with greater literary and narrative complexity.

Keywords : artificial intelligence audiobook production text to speech spoken narration persian language