The proposed research project falls within the field of multilingual language technologies. It constitutes an interdisciplinary research initiative, strongly grounded in recent advances in Natural Language Processing (NLP), whose primary objective is to develop a semi-automated system for multilingual audiovisual content generation aimed at companies, institutions and service providers in the tourism sector seeking to enhance their digital presence and communication across online platforms.

Although NLP technologies such as ChatGPT, Gemini and Copilot have emerged prominently in recent years, their applicability remains limited in specific domains, such as the generation of content for social media platforms within the tourism sector. To address these challenges, the GAMATUR system will incorporate three components based on adapted generative AI models, optimised for the creation of audiovisual and textual content.

The first component will comprise an automatic subtitle generation system based on the transcription of the audio track of videos. It will also support subtitle generation in multiple languages, thereby facilitating wider dissemination and global reach whilst enhancing accessibility. The second component will integrate Large Language Models (LLMs) to enable users to obtain automatically generated descriptions of both entire videos and specific video frames. These descriptions may be translated into any of the languages supported by the system, and synthetic speech may be generated from the textual descriptions. As part of this component, users will also be able to interact with the LLMs through a chatbot, using either written or spoken input. Finally, the system will provide automated audio and video enhancement functionalities based on generative models, together with additional editing capabilities, including automatic video segmentation and cutting, among others.

The design and implementation of these functionalities will be guided by the latest developments in NLP and Artificial Intelligence (AI). Recent advances have demonstrated the effectiveness of LLMs and Transformer-based neural networks, providing the technological foundations required for the development of the proposed system. To ensure the adequate performance of the integrated models and algorithms, these will be fine-tuned and optimised through their combination with comparable corpora and multimodal corpora in the tourism domain, as well as with terminological resources to be compiled throughout the project.

The GAMATUR system will constitute an integrated and advanced technological solution that will not only enhance the efficiency and effectiveness of content creation, but will also generate significant added value for the promotion and development of tourism, particularly in rural areas and specialised sectors such as wine tourism. The system will further facilitate technological application and knowledge transfer, whilst contributing to economic development by strengthening the competitiveness of businesses and the tourism sector in both Andalusia and Spain in relation to other European and global markets, within an increasingly globalised context.