AstraVerse: Designing Culturally-Sensitive Video Game Characters

When gods from Hindu mythology appeared in the popular MOBA game SMITE, the Hindu American Foundation filed a complaint with the FCC. When Age of Mythology included Hindu deities as playable characters, cultural organizations protested the trivialization of sacred figures. These incidents highlight a fundamental tension: game developers want to draw from rich mythological traditions, but when they directly copy religious figures, they risk cultural appropriation and community backlash.

This case study documents the development of AstraVerse, a framework for creating culturally-inspired game characters that are both novel and respectful. Rather than transplanting gods wholesale into games, AstraVerse enables designers to remix cultural elements systematically—creating original characters that honor traditions without copying sacred figures.

The Problem

Mythology and cultural heritage offer game designers a treasure trove of inspiration. Hindu mythology alone contains thousands of deities, each with complex narratives, symbolic attributes, and visual characteristics. But the common industry practice of "god transplantation"—directly copying religious figures into games as playable characters—creates significant ethical, creative, and business challenges.

Why this matters: When gods like Kali or Ganesha appear in combat games, they're often reduced to combat stats and special abilities. Their sacred significance—worshipped by millions—becomes secondary to gameplay mechanics. Cultural communities react strongly: protests, FCC complaints, and boycotts have followed games that misrepresent sacred figures.

This puts designers in an impossible bind: Do they avoid cultural themes entirely and miss out on rich creative opportunities? Or do they risk community backlash by directly copying figures? Neither option serves games, players, or cultural communities well.

Our goal: Develop a systematic framework that allows designers to draw from cultural aesthetics respectfully while creating novel, original characters that don't replicate or misrepresent sacred figures.

My Approach

As Project Lead and Primary Researcher, I developed AstraVerse through a rigorous UX research process combining ethnography, design systems thinking, and participatory validation:

• Conducted visual ethnography at 4 Hindu temples in California, collecting 429 cultural artifacts to understand authentic visual language
• Applied grounded theory analysis to systematically identify 141 granular cultural elements (syntagms) organized into 22 categories (paradigms)
• Developed card-based design toolkit enabling designers to remix cultural elements like building blocks
• Validated through participatory workshops with game designers from UCSC's Game Design program
• Evaluated novelty quantitatively using vector embeddings and statistical analysis
• Assessed cultural sensitivity qualitatively through interviews with Hindu temple priests and cultural experts

Key Outcomes

• 97% novelty rate for created characters
• 100% cultural sensitivity validation from temple priests
• 37 character concepts generated by designers
• Published at ACM FDG 2023 & IFIP-ICEC 2023
• Systematic framework: Card-based design toolkit for cultural inspiration
• Cultural respect: Avoids direct deity copying while maintaining authenticity
• Designer empowerment: Enables creative exploration with cultural guardrails
• Replicable process: Can apply methodology to other cultures and traditions

Visual Ethnography in Cultural Spaces

To understand authentic cultural visual language—not stereotypes or Western representations—we conducted visual ethnography at four Hindu temples in Sunnyvale, California. This wasn't tourism or superficial observation; we systematically documented the actual visual elements, rituals, and objects present in these sacred spaces.

Why temples matter: Hindu temples are living repositories of visual culture where traditions are preserved and practiced. Each idol, ritual object, and architectural detail carries centuries of symbolic meaning. To create culturally grounded characters, we needed to understand these authentic visual vocabularies, not market stereotypes.

Over multiple visits, we collected 429 photographs systematically categorized as:

• 248 Images of Idols: Sculptures of deities showing canonical poses (mudras), clothing styles, ornaments, weapons, and symbolic attributes
• 63 Images of Ongoing Rituals: Live ceremonies revealing contextual uses of objects, gestures, and spatial arrangements
• 45 Images of Miscellaneous Objects: Ritual implements, ceremonial items, architectural elements, and symbolic objects

The methodology: We didn't just photograph and leave. We spent time in each temple, observing multiple worship sessions, speaking with priests about symbolic meanings, and understanding the cultural context of each element. This ethnographic grounding ensured our design framework drew from authentic cultural practice, not surface-level aesthetics.

Understanding Cultural Stakeholder Concerns

Early in the research process, we showed temple priests examples of how Hindu gods appeared in existing games like SMITE and Age of Mythology. These conversations were critical for understanding what cultural appropriation actually looks like from the community's perspective—not just academic critique, but lived experience.

What we learned: The priests' concerns weren't merely about inaccurate representation—they were about sacred figures being reduced to combat stats, sacred narratives being trivialized into game mechanics, and deities worshipped by millions becoming entertainment commodities. These aren't abstract deities in mythology books; they're living figures of devotion.

One priest explained, "When you show Kali fighting other gods, you're showing something that isn't just wrong—it's upsetting to those who worship her." Another clarified the distinction between inspiration and copying: "We respect creative work, but when you directly copy our gods into games, that's different from creating something inspired by our culture."

This insight shaped our framework: Rather than asking "How do we copy gods respectfully?", we asked "How do we create original characters inspired by cultural aesthetics?" This reframing became foundational to AstraVerse.

Phase 2: Building the Design System

Having collected rich ethnographic data, we faced a critical challenge: How do we transform hundreds of photographs of idols, rituals, and objects into a usable design system that enables creativity rather than copying?

The breakthrough: Rather than treating whole gods as units, we identified their component elements—ornaments, postures, symbolic attributes, color palettes. By making these elements modular and remixable, designers could create new combinations while staying culturally grounded.

Grounded Theory Analysis

Using constructivist Grounded Theory, we systematically analyzed the visual data through iterative coding. This wasn't surface-level categorization—we identified patterns, relationships, and structural elements that defined the cultural visual language.

The process: We began with open coding, identifying every visual element present in the ethnographic data: headwear styles, hand gestures, weapons, clothing patterns, symbolic objects, color schemes, postures. Through axial coding, we grouped these into thematic categories. Finally, selective coding revealed the high-level pillars organizing the entire design space.

Phase 3: The AstraVerse Design Toolkit

The framework was translated into a tangible design tool: a set of cards representing the identified cultural elements. This allows designers to search for specific elements they are looking for, similar to how Character Creation Interfaces in games present structured choices.

How the Toolkit Works

• Paradigms help navigate the design space
• Designers select Syntagm cards from various Paradigms
• By combining selected cards, designers can construct new, unique characters rooted in cultural aesthetics but not directly copying existing figures

Phase 4: Validation

A design system is only effective if it meets its goals. We conducted a multi-faceted evaluation focusing on our key research questions.

Evaluation 1: Participatory Design Workshop

Could game designers actually use the AstraVerse system to create original characters? To answer this, we invited three game designers from UCSC's Game Design and Development graduate program to a hands-on workshop.

The task: Designers were given the card-based toolkit and asked to create two characters each, tasked with saving a fictional planet called "Vritra." They had full access to all 141 syntagm cards organized across 22 paradigms, with no restrictions on combinations.

What we observed: Within minutes, designers were actively mixing elements across paradigms. One designer created a character combining a "multi-armed posture" with "trident weapon" and "blue skin tone"—elements from different original gods, but combined into something new. Another paired "lotus mudra" with "armor made of ancient scripts" and "airship mount"—creating a character that felt rooted in cultural aesthetics but was entirely original.

Why this matters: The workshop confirmed that designers could use the framework to generate original characters. But were these characters actually novel, or just remixes that felt new? That's where quantitative evaluation came in.

Evaluation 2: Quantitative Novelty Analysis

Designers might feel like they're creating novel characters, but are they actually different from traditional "god transplantation"? To answer this objectively, we developed a quantitative framework to measure character novelty using vector embeddings.

The method: Each character (both participant-created and original gods) was represented as a 141-dimensional vector based on which syntagms they contained. For example, Shiva might be represented as [1, 0, 1, 0, 1...] where 1 indicates presence of a specific element. Participant characters were similarly encoded based on their selected cards.

We then calculated Canberra distance—a metric accounting for the relative differences between vectors. Lower distances meant characters shared more elements; higher distances indicated greater novelty.

The comparison: We compared participant-generated characters against all original Hindu gods from our ethnographic data. We also broke down comparisons by category (Physical Features, Weapon Design, Fashion, etc.) to understand where novelty manifested.

What this means: The quantitative analysis confirmed that character creation using AstraVerse produced statistically distinct characters from traditional god transplantation. But numbers alone don't prove cultural sensitivity—that required evaluation by cultural experts.

Evaluation 3: Cultural Sensitivity

We interviewed three cultural experts (two Hindu temple priests, one with a diploma in Hindu scriptures). They evaluated participant-generated characters alongside gods transplanted into games.

Key Outcomes & Impact

Conclusion

The AstraVerse project successfully demonstrated that it's possible to move beyond direct appropriation in culturally inspired character design. Through a rigorous UX research and design process, we developed and validated a system that:

• Provides designers with a structured yet flexible tool for creative exploration grounded in cultural aesthetics
• Enables the generation of characters perceived as statistically novel and visually unique
• Produces character concepts deemed culturally sensitive and acceptable by cultural stakeholders