◈ STATEMENT ◈
بيان அறிக்கை

On Human-AI Collaboration

✦ ✦ ✦

The Research

This digital exhibition is the result of scholarly research carried out by human researchers. Every historical argument, every source retrieved and every interpretation presented here is the result of human effort: reading colonial documents, postcolonial studies, and testimonies from those who resisted categorization.

The questions raised by this exhibition, and the answers it offers, belong entirely to its human authors.

The Collaboration

The code for this exhibition was generated using generative AI (Claude, developed by Anthropic). This exhibition was developed in two weeks, and without the help of AI, it would have been difficult for researchers with limited training in web development to create a functional, responsive and multilingual website within this timeframe. We made the choice to use these tools to bridge the gap between scientific vision and technical implementation, while remaining transparent about their use.

We see this more as a collaboration, rather than a delegation. AI worked as a technical assistant that translated human vision into functional code, while the intellectual and creative labor remained in the hands of the researchers.

Human Contributions

  • All historical research and analysis
  • Archival work and source discovery
  • Theoretical framing and argumentation
  • Exhibition narrative and structure
  • Visual aesthetics and design direction
  • Color palette and typography choices
  • Curatorial voice and editorial judgment
  • Multilingual content (Arabic, Tamil, English)

AI Contributions

  • HTML, CSS, JavaScript code generation
  • Technical implementation of design specs
  • Responsive layout and compatibility
  • Code debugging and optimization
  • File format conversions

How It Worked: An Example

Below is an example of how we used AI to generate code. The researcher provided specific design requirements, aesthetic references and content structure. The AI translated these into functional code.

Example Prompt → Code
HUMAN PROMPT:
Create a music player for archival Gallica recordings. It should have: - A dropdown to select tracks (Here are 6 Shikhate Aïta recordings links from BnF [ARK links]) - Display: Arabic title, Latin transliteration, artist, year, source - A "Listen on Gallica" button that opens BnF in new tab - Use the exhibition color palette: parchment dark (#1a1510), ink-gold (#c9a227), morocco-teal (#1a6b6b) - Fonts: Special Elite for labels, Playfair Display for titles - Include Arabic decoration text: صوت العيطة - Make it responsive for mobile
AI OUTPUT:
→ Generated 300+ lines of HTML/CSS/JavaScript → Created dropdown with optgroup organization → Implemented Gallica URL construction → Applied exhibition color variables → Added responsive breakpoints → Human reviewed, tested, and refined

Why Transparency Matters

For an exhibition about how colonial regimes produced knowledge, we believe it is essential to be transparent about how this exhibition itself was produced.

AI tools are not neutral, they are trained using data that contains historical biases, particularly colonial ones. By recognizing the role of AI in our workflow while clearly defining its limitations, we aim to promote the critical and responsible use of these technologies in the field of digital humanities.

✦ ✦ ✦
Ferdaous Affan & Abirami Raveendranath
Digital Global South Fellowship 2025 Centre for Contemporary & Digital History (C²DH) University of Luxembourg
← Return to Exhibition