D2CIN: Advancing Inclusive Education Through Conversational AI

At a glance

D2CIN turns inclusive AI research into practical educational resources. Mimic leads the development of a conversational 3D avatar, supports educator testing in Germany, and co-develops a handbook that helps education and creative professionals build fairer, more accessible AI experiences.

Artificial intelligence is becoming part of how people teach, learn and create. Yet many educators and creative professionals still lack practical ways to identify bias, improve accessibility and apply ethical design principles when building AI-powered experiences.

The Erasmus+ project Digital Diversity: Crafting Inclusive AI Narratives, known as D2CIN, addresses this gap through research, educator engagement, conversational AI and interactive digital characters. Led in partnership with Budakov Films and co-funded by Erasmus+, the initiative is designed to make inclusive AI more understandable, testable and useful in real educational settings.

Mimic’s role is to translate the project research into an experience people can actively use. At the centre of that work is the Personalized Feedback Navigator, a conversational 3D avatar that provides reflective, context-aware guidance on inclusive AI design.

Project Snapshot

Field	Details
Project	Digital Diversity: Crafting Inclusive AI Narratives (D2CIN)
Release	September 2025, with a 12-month project duration
Format	Educational initiative / Research and development
Audience	Education providers, educators, freelancers and creative practitioners
Services	Conversational AI, digital avatars and interactive storytelling
Core outputs	Personalized Feedback Navigator, Accessible AI for Education Handbook, Inclusive AI Impact Pilot and outreach events
Partners	Mimic Productions and Budakov Films
Funding	Co-funded by Erasmus+
Platform	Academic and educational use

D2CIN addresses a practical gap in AI education. Many educators and creators understand that bias, accessibility and ethical design matter, but they may not have the tools, technical confidence or structured guidance needed to apply those principles in real projects.

The project combines research with usable resources. Rather than treating inclusive AI as a purely theoretical subject, D2CIN gives participants concrete methods for evaluating conversational systems, designing more accessible content and reflecting on how cultural context shapes user experience.

Conversational learning is especially valuable in this context. A handbook can explain principles, while an interactive guide can ask questions, respond to uncertainty and help users work through decisions. The avatar format makes the process more active, approachable and memorable without presenting itself as a replacement for educators or human expertise.

What the Pilot Research Is Revealing

Focus groups and workshops in Bulgaria and Germany have generated early insights that are shaping the project tools, educational resources and avatar experience.

Interaction outperforms passive content

Participants engage more deeply when they can question a character, consider choices and receive feedback. These interaction loops can strengthen attention, retention and emotional connection compared with one-way content.

Context matters as much as neutrality

Avoiding bias is essential, but an AI character must also understand cultural and situational context. A system that treats every user and scenario as identical can become less helpful, not more inclusive.

Trust depends on transparency

Users respond positively when an avatar is predictable, clear about its role and non-judgmental. Authentic, consistent behaviour matters more than exaggerated friendliness or attempts to appear overly human.

Personalisation needs boundaries

Adaptive guidance can improve learning, but excessive personalisation may introduce new bias, confusion or discomfort. User control, explainability and restraint are therefore treated as core design requirements.

There is no single ideal visual style

Some users find realistic avatars relatable, while others feel more comfortable with stylised or non-human characters. The right choice depends on the audience, context and purpose of the interaction.

Inclusive AI is not achieved by making every interaction the same. It comes from designing systems that are fair, transparent, context-aware and controllable by the people using them.

The Personalized Feedback Navigator

Following two focus groups in Bulgaria and Germany, the project team began consolidating observations, expert recommendations and participant reflections into the Personalized Feedback Navigator.

Delivered as an interactive 3D avatar, the Navigator helps educators and freelancers examine how inclusive and bias-aware their conversational AI solutions are. Through conversation, it provides reflective feedback on accessibility, ethical use, content design and user experience.

The goal is not to issue a simple pass or fail result. It is to help users identify gaps, consider alternative perspectives and make informed improvements within their own working context.

The avatar creates a more approachable interface for complex guidance. Users can explore recommendations through dialogue, clarify unfamiliar concepts and move through the information at their own pace. This transforms assessment into a guided learning experience while keeping the user in control.

The Personalized Feedback Navigator is available through the project website and represents an important step toward more accessible, human-centred interfaces for knowledge transfer and professional development.

Exploring Inclusive AI with Students and Educators in Berlin

As part of the project, Mimic hosted two workshops in Berlin that brought together students, educators and AI practitioners. The sessions combined discussion with hands-on testing, enabling participants to explore inclusive AI principles through real learning scenarios rather than abstract examples.

Participants worked with project resources, tested conversational and avatar-based tools, and discussed how digital humans could support communication, accessibility and learner engagement. The workshops also created space for critical reflection on where AI guidance is useful, where it may cause discomfort, and how educators can retain control of the learning process.

A key part of the sessions was direct interaction with conversational AI avatars. This allowed participants to evaluate not only what the systems communicated, but also how character design, tone, visual style and responsiveness influenced trust and comprehension.

The feedback gathered in Berlin contributes directly to the continued development of the Navigator, the handbook and the project’s wider recommendations for inclusive AI in education.

Mimic’s Contribution

Conversational AI Research and Development

Mimic leads the technical development of the Personalized Feedback Navigator, drawing on its experience in conversational AI, real-time digital characters and interactive systems. The avatar is designed as an ethical guide that supports reflection rather than replacing human expertise.

Digital Avatar Experience Design

The project requires more than a chatbot placed inside a character. The avatar’s visual presence, conversational behaviour, feedback structure and interaction flow must work together to create a credible educational experience. Mimic applies digital-human craft to make the guidance clear, engaging and suitable for real-time use.

Educational Outreach and Validation

Mimic works with education providers and subject experts to test project tools in practical settings. The project aims to engage 150 educators in hands-on activities, helping validate the resources and bridge the gap between industry innovation and educational practice.

Strategic Project Leadership

Working alongside Budakov Films, Mimic co-develops the Accessible AI for Education Handbook and contributes to the project’s wider framework for digital diversity. This ensures that technical development is supported by practical, ethical guidance for the vocational education community.

Project Outputs

Personalized Feedback Navigator

An online conversational avatar that helps users assess inclusivity, accessibility and bias awareness in AI-driven content and interactions.

Accessible AI for Education Handbook

A practical guide covering bias reduction, accessibility, ethical AI use and inclusive user experience through clear steps and applied examples.

Inclusive AI Impact Pilot and Outreach Events

Workshops and launch activities in Germany and Bulgaria where participants test the tools, provide feedback and share inclusive AI practices across Europe.

Why Conversational AI Supports Inclusive Learning

Conversational AI can place learners inside an active process. Instead of only receiving information, users can ask questions, revisit assumptions, explain their reasoning and receive feedback while the topic is still relevant.

AI avatars add social presence to this model. A well-designed digital guide can make complex material easier to approach, adapt the pace of explanation and create a clearer interaction structure. For sensitive topics such as bias, inclusion and accessibility, a non-judgmental interface may also help users explore uncertainty more openly.

The D2CIN findings also show why restraint matters. The character should not manipulate, over-personalise or claim authority it does not have. A responsible educational avatar is transparent about its purpose, grounded in structured guidance and designed to preserve learner agency.

Shaping More Inclusive Digital Learning

D2CIN demonstrates how digital characters can move beyond novelty and become practical interfaces for education. By combining conversational guidance with research-backed resources, the project makes inclusive AI principles easier to understand, evaluate and apply.

The wider opportunity extends across teacher training, professional development, scenario-based learning, cultural interpretation and accessibility guidance. Conversational avatars can also make institutional knowledge more interactive, particularly when learners need to ask questions or reflect on decisions rather than simply read information.

For Mimic, the project reflects a broader vision for embodied AI. Digital humans create value when premium character design, responsible conversational behaviour and a clear use case are developed together. In education, that means creating experiences that are engaging without becoming distracting, personalised without becoming intrusive, and intelligent without losing human oversight.

Credits

Project	Digital Diversity: Crafting Inclusive AI Narratives (D2CIN)
Funding	Erasmus+
Creative Partner	Budakov Films Ltd
Mimic Role	Strategic project leadership, conversational AI avatar R&D, ethical AI development and workshop hosting
Audience	Education providers, educators, freelancers and creative practitioners
Technology Focus	Conversational AI, 3D digital avatars and interactive storytelling

Explore the Project

Learn more about D2CIN, access the project resources and explore the Personalized Feedback Navigator at https://d2cin.eu/.