AI-powered digital games with stealth assessment and creativity support tools can effectively measure and enhance human creativity while promoting deeper learning in educational contexts.
Objective: The main goal of this research was to explore how AI-powered digital games can be used to assess and support creativity in educational settings. Specifically, the study aimed to investigate the effectiveness of different creativity support systems in digital games, develop stealth assessment methods for measuring creativity without human judgment, and establish a foundation for integrating creativity into game-based learning environments to promote deeper understanding of academic subjects.
Methods: The research employed multiple methodological approaches across several interconnected studies. The primary experimental study used a pretest-posttest design with 114 college students randomly assigned to four conditions while creating game levels in Physics Playground: (1) Inspirational condition with example levels, random word prompts, and brainstorming tools; (2) Instructional condition with structured three-step design process; (3) Both conditions combined; and (4) Control condition with no creativity supports. Creativity was assessed through external measures including the Alternative Uses Test and figural creativity tasks inspired by the Torrance Test of Creative Thinking, as well as through consensual assessment of student-created game levels. Additionally, the research implemented stealth assessment using Evidence-Centered Design framework with Bayesian network modeling to continuously capture and analyze student interaction data during gameplay, identifying observable behaviors that indicate creativity competencies like fluency, flexibility, and originality.
Key Findings: The research revealed several significant findings about creativity support and assessment in digital game environments. Students in the Inspirational condition created significantly more creative game levels compared to those in the Instructional and No Supports conditions, while the combination of both inspirational and instructional supports showed the most promise for enhancing creativity. The stealth assessment system successfully demonstrated convergent validity, with significant correlations between AI-powered creativity estimates and traditional external creativity measures like the Alternative Uses Test. The study found that most creativity enhancement came from inspirational supports rather than instructional guidance, suggesting that exposure to examples and idea-generation prompts are more effective than structured step-by-step instructions. The research also established that digital games can serve as effective platforms for both assessing and cultivating creativity through features like creative gameplay mechanics and content creation tools.
Implications: These findings have substantial implications for AI in education, particularly in developing intelligent tutoring systems that can dynamically assess and support student creativity. The successful validation of stealth assessment techniques opens possibilities for real-time, personalized creativity support in educational games without interrupting the learning flow. The research demonstrates that AI can augment human creativity rather than replace it, supporting the concept of Intelligence Augmentation over pure Artificial Intelligence. The integration of creativity assessment with subject matter learning suggests new pathways for developing educational games that simultaneously teach academic content and enhance creative thinking skills. The work provides a foundation for creating adaptive learning environments that can identify individual creativity profiles and provide targeted support based on specific creative competencies like flexibility, fluency, or originality.
Limitations: The study acknowledges several limitations that affect the generalizability of findings. The experimental study involved a relatively small sample size of 114 college students, which may not represent broader populations or different age groups. The research focused primarily on one specific digital game environment (Physics Playground), limiting the applicability to other educational contexts or subject domains. The stealth assessment validation, while promising, requires additional validation studies across different domains and populations before widespread implementation. The creativity assessment relied heavily on consensual assessment techniques, which may introduce subjective biases despite efforts to maintain objectivity. Additionally, the study's focus on short-term creativity outcomes may not capture long-term creative development or transfer effects to other domains.
Future Directions: The research outlines ambitious future directions for advancing AI-powered creativity assessment and support. Plans include expanding stealth assessment validation to other domains like computer science and music through platforms like EarSketch, where students write code and create music remixes. The author proposes leveraging machine learning to identify new patterns in user behavior data that weren't captured in traditional top-down assessment models, potentially improving assessment accuracy through bottom-up approaches. Future work will explore AI-powered Natural Language Processing for analyzing written and spoken creativity, as well as computer vision techniques for assessing visual creativity in drawings and videos. The research agenda includes developing Creative Game-Based Learning (CGBL) frameworks that use creativity to deepen subject matter learning, with Physics Playground 2 serving as a testbed for this approach. Long-term goals involve creating comprehensive digital environments that can accurately assess multiple creativity facets and provide personalized, real-time creativity support tailored to individual needs and learning contexts.
Title and Authors: "Going Beyond the Brick: Assessing and Supporting Creativity Using AI-Powered Digital Games" by Seyedahmad Rahimi.
Published On: August 7, 2023
Published By: Creativity Research Journal