Eliciting Prompt Responses: the Role of Enhanced GenAI Syntax in Image Generation Tools. Using GenCraft to Develop Transversal Skills in Learning Environments

Artificial Intelligence (AI) has represented a revolutionary breakthrough, encompassing a plethora of human-related activities and services (Huang / Rust 2021; Curry et al. 2024). Recently, AI has been boosted by the development of LLMs (Large Language Models), giving rise to the Generative AI paradigm (GenAI), which “focuses on developing algorithms and models capable of generating synthetic data that closely resemble real-world data” (Bandi et al. 2023: 1). From a linguistic (and transdisciplinary) perspective, GenAI has the potential to create any contents via human-induced prompts. Along with text-to-text GenAI, other services aim at generating realistic, ready-to-use contents. AI-based image generators are a case in point. Based on a simple textual prompt, such services are able to create realistic representations from scratch. Against this background, this investigation focuses on the importance of textual prompts in making the most of the potential offered by GenAI image generators; in particular, the analysis takes GenCraft, a freemium app and website, into consideration. Here, the text box used to describe the image-to-be embeds a button called Enhance Prompt, an AI-based tool aimed at fine-tuning the suggested initial text by users. Such feature represents an interesting and challenging mechanism that re-defines language patterns, characterised by a wise combination of technical syntax and semantic enrichment. An on-field assessment of such generative language outputs and a comparison of human-induced vs. AI-based prompts in creating perfected images demonstrates the synergy between AI and human interaction aimed at enhanced results and tasks for the development of mixed abilities, towards a wise and sensitive use of AI to be advocated in learning-related environments.