Educational games enhanced with generative AI technologies offer promising opportunities to improve writing skills across the curriculum for students with disabilities, providing personalized learning experiences that increase engagement and academic achievement.

Objective: The main goal of this study was to explore how game-based learning and generative AI can be integrated to enhance writing across the curriculum, particularly for students with disabilities, while identifying key features educators and families want in educational games.

Methods: The researchers conducted focus groups with 21 participants (17 teachers and 6 parents/caregivers, with some overlap) to gather perspectives on desired features in educational games. Participants were primarily female (19 females, 2 males) with ages ranging from 28 to 65 years. The teachers represented elementary, middle, and high schools across public and private settings in four states, with an average of 15.5 years of teaching experience. Data was collected through five separate focus group sessions conducted over Zoom, each lasting approximately 70 minutes. The researchers followed a semi-structured interview protocol, analyzed transcripts using a constant-comparative method of qualitative analysis, and used triangulation to ensure credibility and trustworthiness of findings.

Key Findings:

• Participants identified five key features for effective educational games: (1) content-driven and relevant to curriculum objectives; (2) accessible and easily navigable for students with disabilities; (3) personalized to allow for character customization and adaptable difficulty levels; (4) collaborative, engaging, and interactive; and (5) capable of providing meaningful and usable data for teachers and parents.
• The researchers used these insights to develop prototype games featuring "Writing Warriors" (animal superheroes) that help students with different aspects of the writing process, including brainstorming, drafting, and editing.
• Games incorporated adaptive leveling (automatically adjusting difficulty based on student performance), immediate feedback, and reward systems to maintain engagement.
• The WEGO tool (Writing Efficiently with Graphic Organizers), an existing technology-based writing intervention, was enhanced with AI capabilities to support the brainstorming phase, an area where students consistently struggled.
• Teachers emphasized the need for "guardrails" around AI features to prevent issues with content and bias, as well as the ability to enable or disable AI support based on individual student needs.

Implications: The findings contribute to the field of AI in education by:

• Demonstrating how educational games with AI support can help learners with disabilities develop writing skills across the curriculum, including STEM contexts.
• Providing a framework for designing educational games that are accessible, engaging, and effective for diverse learners.
• Highlighting the potential of AI to personalize learning experiences, provide immediate feedback, and generate customized content.
• Showing how teacher control and customization of AI features are essential for successful implementation in educational settings.
• Offering a practical example of how AI can be integrated into existing evidence-based interventions to address specific learning challenges (such as brainstorming in writing).

Limitations:

• The gender imbalance in participants (19 females, 2 males) limits the generalizability of findings, as male perspectives, particularly from fathers, were underrepresented.
• The study represents an early exploratory phase of design research rather than a completed implementation with effectiveness data.
• Generative AI tools can produce "hallucinations" and errors in STEM content, potentially reinforcing misconceptions.
• AI tools may contain biases, raise privacy and data security concerns, and lack accessibility features needed for students with disabilities.
• The prototype games and AI enhancements are still in conceptual stages, with ongoing development needed before widespread implementation.

Future Directions:

• Continue developing and refining the prototype games with AI features based on user feedback.
• Implement and evaluate the effectiveness of the AI-enhanced WEGO tool in diverse educational settings.
• Explore how generative AI can support other aspects of writing beyond brainstorming.
• Investigate ways to minimize AI biases and ensure data privacy and security for students with disabilities.
• Develop AI literacy curricula for students with disabilities to help them critically evaluate AI-generated content.
• Expand research on using AI to support STEM writing activities for students with diverse learning needs.

Title and Authors: "Educational Games and the Potential of AI to Transform Writing Across the Curriculum" by Anya S. Evmenova, Kelley Regan, Reagan Mergen, and Roba Hrisseh.

Published On: May 2, 2025

Published By: Education Sciences (Volume 15, Issue 5)