BioVrium
VR Platform
Enhancing Microbiology Education with VR and AI
Project type: Educational Technology Innovation
Role: UX Researcher, UX Designer, VR Developer, Data Analyst
Industry: Educational Technology
Tools: Unreal Engine 5.3.2, CONVAI, SUS, InfraNodus
Duration: Research and development phase over several months
Introduction
The users
Participants included students and individuals interested in VR and science education, aged 25-55, with diverse educational backgrounds.
How can we improve learning effectiveness?
Challenges Identified
Complex Subject Matter: Microbiology is inherently complex, requiring a thoughtful approach to make content accessible and engaging.
Diverse Learning Styles: Catering to different learning preferences and ensuring inclusivity.
Technological Adoption: Ensuring users are comfortable with VR technology and AI interactions.
Approach
User-Centered Design: Designing the platform based on the identified needs and preferences from user research. This included realistic 3D models, interactive content, and built-in quizzes for progress tracking.
Information Architecture: Structuring content into four progressive levels: Virtual Classroom, Inside the Microscope, Virus World, and Bacteria World. This organization helps break down complex information into manageable chunks.
Interaction Design: Developing intuitive VR controls for manipulating 3D models and voice interaction with AI characters to create a more immersive experience.
AI Character Integration: Creating three distinct AI characters (BioBear, Anca, Dr. Ellis) to cater to different learning styles and preferences, enhancing user engagement.
Accessibility Considerations: Designing with inclusivity in mind, ensuring the platform is accessible to users with varying needs and preferences.
UserFlow of User’s interaction with BioVrium
THE PROPOSED SOLUTION
To investigated weather the presence of AI can truly make an impact on user engagement and learning outcomes, I created 2 versions of BioVrium: one with the AI characters and one without. The content is structured into four progressive levels: Virtual Classroom, Inside the Microscope, Virus World, and Bacteria World, allowing students to explore detailed 3D models and simulations. The platform features intuitive VR controls, voice interactions with AI characters, and built-in quizzes with immediate feedback to enhance user engagement. Three distinct AI characters (BioBear, Anca, and Dr. Ellis) cater to different learning styles and preferences, while accessibility and inclusivity are prioritized through adaptive learning and multilingual support. Supplementary learning materials and progress tracking ensure a comprehensive learning experience.
Level 1 - The Classroom
Goal
To assess the usability and effectiveness of BioVRium, comparing the AI-enhanced version to a VR-only version.
Findings
User research is fundamental to make informed design decisions, as it ensures the platform aligns with the needs and preferences of its users. The integration of AI has shown significant potential in enhancing learning experiences by providing personalized and interactive support. Additionally, the diversity in AI character design proved to be valuable, as it catered to various learning styles and preferences, thereby improving user engagement and satisfaction. For future improvements, enhancing AI performance, adding multilingual support, and incorporating additional features will be crucial steps to further elevate the user experience and accessibility of the platform.
Despite the limitations encountered during this study, the findings were statistically significant and should be further investigated. Incorporating generative AI in immersive Virtual Reality platforms for education can lead to higher engagement, improved usability and academic performance. However, the full potential of generative AI remains to be explored.
What I learned
Study 1 - Understanding the Users’ Needs and Challenges
Goal
To identify user needs, preferences, and pain points related to microbiology learning and VR usage.
Methodology
An online survey distributed through a professional VR research community completed by 31 participants.
Level 2 - Inside the Microscope
BioBear
an animated teddy bear, is designed to transform learning into a playful and engaging experience. His presence in the VR environment is intended to reduce the intimidation often associated with complex scientific subjects like microbiology. BioBear's approachable and friendly nature makes it particularly appealing to users who might find traditional educational settings daunting or who respond well to gamified learning environments.
Anca
a young, knowledgeable microbiologist who combines expertise with a relatable and friendly demeanour. Her character is intended to bridge the gap between rigorous academic knowledge and modern, interactive learning methods. Anca is particularly suited to learners who seek a balance between credible scientific figure and a supportive, engaging learning experience.
Dr. Ellis
the archetype of a traditional academic expert. Portrayed as a senior, knowledgeable professor Dr. Ellis appeals to learners who prefer a more structured and authoritative approach to education. His persona is crafted to reflect deep expertise and reliability, aiming to attract users who appreciate thorough explanations and a serious educational tone.
Study 2 - Usability Evaluation with a crossover design
Key
Insights
Methodology
A crossover study with 10 participants involving pre- and post-assessments, UX questionnaires (including SUS), observational notes, and audio recordings.
Thematic Analysis for qualitative data
Key
Insights
Target Audience
Students and individuals interested in VR and/or science education.
-
When asked about the most essential features of a VR educational platform for microbiology, the majority - 54.8% chose 'All of the above,' emphasizing the need for realistic 3D models, interactive content, and adaptive learning paths, which were subsequently incorporated into the platform.
-
The vast majority (32,3%) rated built-in quizzes and assessments as 'Very important', leading to their integration in both the AI-enhanced and control versions of the platform.
-
When asked about the most essential features of a VR educational platform for microbiology, the majority - 54.8% chose 'All of the above,' emphasizing the need for realistic 3D models, interactive content, and adaptive learning paths, which were subsequently incorporated into the platform.
-
In the development BioVRium, the decision regarding the session’s duration was informed by both current literature and user feedback. Sumardani and Lin’s (2023) study suggested that the optimal duration for virtual reality learning is less than 15 minutes. However, the participants indicated a preference for longer sessions, with an average suggestion of 30 minutes. Taking both these factors into account, the sessions last 20-25 minutes.
The Product
BioVRium is an immersive VR learning platform enhanced with generative AI. It aims to overcome the limitations of traditional microbiology education by allowing students to interact with detailed 3D models of microorganisms in a dynamic virtual environment.
Level 3 - Inside the Virus Wolrd
What did not work?
“Without AI because I did not need to wait for a long period to know what the answer to the question asked is.”
〰️
〰️
While many participants enjoyed the interaction with AI guides and found them helpful, some faced challenges that impacted their experience. P1 and P8 appreciated the fun and ease brought by AI guides like Biobear and Anca. However, P3 found it inconvenient to return to specific locations to interact with the AI, and both P5 and P6 mentioned delays in the AI's responses. These issues were attributed to the CONVAI plugin, which had limitations in real-time responsiveness. Despite these challenges, participants saw the potential benefits of AI, suggesting that resolving these technical issues could significantly enhance the learning experience.
Target Audience
Students of University of West London studying related and non-related fields.
-
Pre- and post-session assessments were utilized to measure knowledge acquisition directly which is a valid tool for assessment. The quantitative data allowed for a comparative analysis to determine if participants learned more effectively with the AI-enhanced VR system. Additionally, interaction logs from the AI system offered deeper insights into the areas where participants sought further information, which helps in identifying aspects of the platform that could be most beneficial or would need enhancement.
-
User experience and usability were evaluated through a combination of the System Usability Scale (SUS) and detailed qualitative feedback gathered via the UX form and observational data during the sessions. This part of the study specifically examined whether AI integration positively impacts the usability of the VR system and motivates learners more than when using VR alone.
-
By analysing session logs and interaction data, it was assessed whether the inclusion of AI elements influenced participants’ willingness and consistency in engaging with the VR content. Furthermore, it was possible to track specific behaviours such as the frequency of attempts participants made to interact successfully with 3D models, the time taken to complete designated tasks, and their overall navigational efficiency in the VR space.
-
Almost all the participants preferred the platform with the AI characters integrated. The majority of participants preferred Anca, the character that in their opinion, is the most humanized. However, Anca was the only character with a non-human voice. This suggests that users might appreciate certain human elements, such as physical aspect, rather than a humanized voice.
