The foundation of proficient language use, especially in the context of learning English as a foreign language (EFL), is a robust vocabulary. The size of a student’s vocabulary significantly impacts their reading comprehension, communication abilities, and overall academic success. Research suggests that students with extensive vocabularies are better readers (Quines, 2023), more effective communicators (KILIÇ, 2019), and have a deeper understanding of the texts they encounter. Therefore, the importance of increasing students’ vocabulary size cannot be overlooked, as it is intrinsically associated with their success in both academic and real-world settings. To expand students’ vocabulary volume, the integration of Artificial Intelligence (AI) into educational practices has proved effective in enriching the vocabulary learning experience. One such approach is the utilization of AI to generate images that facilitate vocabulary learning. This article will explore how XIPU AI can be applied to enhance vocabulary learning through practical examples and evaluate its benefits and concerns.

 

To review the vocabularies learned in previous classes, instructors can task students with generating an image individually. This exercise facilitates the review of vocabulary knowledge by encouraging students to apply correct spelling, grammar, and sentence structure while recalling as many words as possible in a creative context. By creating an image that accurately represents the words they have learned, students can further internalize these linguistic elements and improve vocabulary retention. The following scenario outlines the steps involved in generating an image to reinforce vocabulary acquisition.

Instructors need to demonstrate how to write an effective prompt in class. First, start a prompt with an actionable verb and the theme of the targeted image (e.g., Generate an image of/Visualize a/Imagine a teacher’s office). The second step is to add specific description details, such as objects, colors, position, and background, to enrich the prompt. In this process, students are required to customize the prompts using their own vocabulary knowledge and creativity. For example, the demonstration prompt visualizes a teacher’s office by describing specific office items, object colors and their positions (Image 1). This not only reinforces vocabulary such as "landline phone," "stationeries," and "wing chair" learned in the previous lesson, but also makes the learning experience more engaging and enjoyable. Once students are familiar with the prompting process, this review task becomes more efficient over time.
 

 

Image 1: Sample prompt demonstrated to students

 

Image 2: Sample image generated by XIPU AI

 

After analyzing the applied prompt, XIPU AI generates an image in approximately 30-40 seconds. Students are encouraged to adapt their prompts regarding theme objects, item color, or position to generate multiple versions of images, thereby allowing them to choose the one that best fits their needs. It has been observed that students who actively refine their prompts often demonstrated better retention of vocabulary items in subsequent tasks.

After students finalize their images, they are instructed to submit them on Padlet. Instructors can then showcase the results to the entire class, either by complimenting individual images containing the most vocabulary items or acknowledging the efforts of all students. Below are some in-class practice images generated by students. 
 

 

Image 3: A Teacher’s Office

 

 

Image 4: A Greengrocer

XIPU AI can also be leveraged to develop interactive learning activities that reinforce vocabulary acquisition. Educators can use AI-generated images to create engaging activities such as "matching images with words or definitions" and "describing and guessing words from images". Instructors can also generate images that prompt students to work in pairs, discussing observed differences. This activity not only enhances vocabulary comprehension and pronunciation practice but also makes the learning process more enjoyable. The following scenario outlines the steps involved in generating an image for a "Find the Differences" speaking task on the topic of "Education".

Before class, instructors can generate two similar images using XIPU AI. When writing the prompt, it is important to specify some significant differences between the two images. For instance, the prompt might be: “Generate two images of the same background and the same angle about a teacher’s office. On the first image: there is a work desk, on which there is a desktop computer, a landline phone, a pen container and a mug. On the second image: there is a work desk, on which there is a desktop computer, a cell phone, a pen container and a calendar”. 

 

Image 5: Example image of “Find the Differences”

 

To gain insights into students' perceptions of teaching activities involving AI image generation for vocabulary enhancement, a survey was conducted with 72 participants from three first-year EAP groups. The survey comprised six questions focusing on the effectiveness of using AI-generated images for vocabulary review and the challenges encountered during the image generation process. The results indicate that students generally have a positive attitude towards incorporating AI images into vocabulary teaching activities.

First, students confirmed the effectiveness of AI-generated images assisting vocabulary retention. In the attitude test, participants rated on a five-point Likert scale, ranging from ‘not effective’ to ‘very effective’, to reflect their perspectives on the effectiveness of utilizing AI-generated images for vocabulary learning. Figure 1 shows a mean score of M = 3.75 (SD = 0.968 N = 72), suggesting a generally positive attitude. Notably, a significant majority (69.44%) deemed this method as either effective (50%) or highly effective (19.44%) for vocabulary review. Moreover, students frequently described the learning method as “fun”, “vivid”, “interesting”, “visualizing words”, “interactive”, “practical”, “meaningful”, “immersive”, “is easier to memorize new words” and “effective since generated images match with the learned words”. Many highlighted its practicality, meaningfulness, and immersive nature, noting it made memorizing new words easier. In addition, most participants (64.5%) claimed that generating images with XIPU AI motivate them to engage more with vocabulary use, thereby leading to enhanced vocabulary retention. This heightened motivation transformed into higher engagement, resulted in repeated attempts to generate images with AI (see Fig. 2). Over 40% of students attempted twice to achieve satisfactory outcomes for target vocabulary items, while over one-third experimented three to five times, further familiarizing them with word spelling and meaning. 

 

 Fig. 1. Students’ perceptions of the effectiveness of using AI-generated images for vocabulary learning

 

 Fig. 2. The number of times students tried to finalize their images

Despite its benefits, students also pointed out the challenges of this practice. First, one primary issue is that XIPU AI cannot display all specific content in a prompt. One student commented that “When I give some words, the AI can’t generate the accurate image for all the factors”. Another reported similar problem, “They cannot really catch all the requirements I put in for generating the images. When I try to put in all the elements, the images generated by AI are always unsatisfying and weird. Sometimes when I put in less* words, the result is better. I don’t know how to balance”. Therefore, many students describe the AI as “silly” or “stupid”. Another challenge is insufficient or inaccurate description from student themselves.  One participant mentioned, “I didn’t use AI frequently, and I still adapt to using AI in my daily study”. Similarly, another added, “Sometimes I could not describe it accurately, so the image might not be ideal for me”. Additionally, overly detailed prompts sometimes resulted in images that defied common sense. Technical issues are another concern. As one student noted, “sometimes it is forgetful”, a prompt usually works once and conversional instructions of adding extra details to the original image does not work well. Another issue is that sometimes there are errors and “it even can’t generate an image” after submitting prompts. 

 

Using XIPU AI to generate images for vocabulary teaching offers numerous advantages for teachers. First, the primary benefit is the significant reduction in class preparation time. According to Callow (2012), it can be labor-intensive and time-consuming to find appropriate and engaging images manually. With XIPU AI, educators can quickly produce high-quality visual content that aligns with their lesson plans. Moreover, the customization options allow teachers to tailor images to specific contexts, incorporating relevant symbols and terminology that match the curriculum. This not only ensures these images are directly relevant to the target words, but also makes the learning materials more intriguing. By streamlining this process, teachers can concentrate more on designing teaching activities and less on administrative tasks, ultimately improving class preparation efficiency.

For students, the integration of XIPU AI in vocabulary learning introduces a more engaging way to consolidate vocabulary retention. Visual aids generated with AI tools establish a strong visual-cognitive connection, allowing brain to interpret these visual representations both perceptually and cognitively, which helps students better remember the meaning of the targeted vocabulary (Vu et al., 2022). This approach transforms potentially mundane memorization task into an interactive and enjoyable learning experience, as students are more likely to relax and engage in classes that emphasize visual content over traditional lectures (Barry, 2002). Moreover, students can participate in various activities such as matching games, story creation, or discussion prompts using these images, making the vocabulary acquisition process more intuitive and fun. Additionally, the personalized nature of these visual aids ensures that the content is directly relevant to their studies, enhancing their interest and motivation to learn (Callow, 2023). For example, if an instructor guides students to review vocabularies under the topic of Healthy Diet, the quality of a generated image depends on correct word spelling and the number of relevant vocabularies recalled by the student. As a result, students are likely to find vocabulary lessons more relatable and interesting, leading to improved comprehension and retention.

 

While XIPU AI excels at generating images, valid concerns remain about the quality and accuracy of these visuals. The AI-generated images may inaccurately depict the number, position, or colors of objects, as well as the spelling of vocabularies. Such inaccurate representations can mislead students and compromise the integrity of educational content, therefore, necessitating instructors’ review and validation on the images before incorporating them into lessons to ensure their effectiveness (Rubman, 2024). 

Additionally, the use of AI in education raises ethical considerations regarding data privacy and intellectual property. Instructors should be aware of how their input data is used and stored by AI tools, and questions about ownership and rights to AI-generated content need to be addressed to ensure compliance with legal standards and ethical guidelines (Burrus, Curtis, & Herman, 2024). It is recommended that AI-generated images should be clearly labeled or annotated by its creators. 

 

In conclusion, the integration of AI-generated images in vocabulary learning represents a forward-thinking approach that aligns with contemporary educational practices. This method not only augments students’ vocabulary retention and spelling but also introduces an element of fun that can significantly boost their motivation and enthusiasm for learning. Despite some concerns, embracing such technological advancements can substantially enhance teaching methodologies, ultimately leading to more effective and engaging learning experiences. Educators are encouraged to explore these tools, adapting and integrating them into their teaching strategies to provide a more dynamic, interactive, and enjoyable learning environment that promotes vocabulary expansion and retention.

 

 

 

References:
Barry, A. M. (2002). Perception and visual communication theory. Journal of Visual Literacy, 22(1), 91-106.
Burrus, O., Curtis, A., & Herman, L. (2024). Unmasking AI: Informing Authenticity Decisions by Labeling AI-Generated Content. ACM Interactions, 31(4), 38-45. https://dl.acm.org/doi/abs/10.1145/366532
Callow, J. (2012). The Rules of Visual Engagement: Images as Tools for Learning. Screen Education, 65, 72-73. https://www.researchgate.net/publication/281069694_The_Rules_of_Visual_Engagement_Images_as_Tools_for_Learning
Callow, J. (2023). Unlocking Memory with AI: The Impact of Image-Assisted Vocabulary Learning on Student Engagement and Motivation. Journal of Educational Technology and Learning, 25(2), 78-89. https://www.researchgate.net/publication/345678901_Unlocking_Memory_with_AI_The_Impact_of_Image-Assisted_Vocabulary_Learning_on_Student_Engagement_and_Motivation
KILIÇ, M. (2019). Vocabulary Knowledge as a Predictor of Performance in Writing and Speaking: A Case of Turkish EFL Learners. PASAA, 57, 134-150. https://files.eric.ed.gov/fulltext/EJ1224421.pdf 
Quines, Z. M. (2023). The impact of students’ vocabulary level to their reading and writing performance. International Journal of English Language and Linguistics Research, 11(2), 18-32. https://doi.org/10.37745/ijellr.13/vol11n21832
Rubman, J. (2024). Supporting Learning with AI-Generated Images: A Research-Backed Guide. MIT Sloan Teaching & Learning Technologies. https://mitsloanedtech.mit.edu/2024/03/06/supporting-learning-with-ai-generated-images-a-research-backed-guide/
Vu, N. N., Hung, B. P., Van, N. T. T., & Lien, N. T. H. (2022). Theoretical and instructional aspects of using multimedia resources in language education: A cognitive view. Multimedia Technologies in the Internet of Things Environment. 2, 165-194.