Bangladesh's copyright law silent on AI authorship questions
Bangladesh's Copyright Act of 2023 modernised digital protections but left a critical gap: it says almost nothing about who owns works created by artificial intelligence or whether those works deserve copyright protection at all.
Tools like ChatGPT, Gemini, and Midjourney now generate text, images, and code with minimal human input. This capability breaks a foundational assumption built into copyright law for centuries-that creative works come from human intellectual effort.
The philosophical problem
Copyright law rests on two main ideas. The first, traced to philosopher John Locke, holds that people own what they create through their own labour and skill. The second, from Immanuel Kant, treats creative works as extensions of the author's personality and identity.
AI-generated works destabilise both concepts. When a machine produces expressive content with no meaningful human creativity involved, copyright law cannot easily answer: who is the author, and what exactly is being protected?
A growing practical problem
The issue extends beyond theory. A 2025 investigation found that AI models generated realistic copies of sensitive documents-including modified national ID images-without flagging data protection concerns. Bangladesh's current copyright framework, designed around human authorship, offers no clear way to address such outputs.
Copyright law historically depends on tracing expressive content back to identifiable human intellectual activity. Modern AI increasingly exposes the weakness in assigning authorship where genuine human creativity may not exist.
The concern is no longer just about copying. AI systems now imitate style, tone, structure, and expressive identity itself-capabilities once associated only with human creators.
A practical path forward
Bangladesh does not need overly complex rules that try to scientifically determine whether every work is human-created. Instead, the law could require creators, publishers, and copyright applicants to disclose whether generative AI tools played a substantial role in production.
Courts could then assess disputes case-by-case by examining factors like the level of human creativity, editing, judgment, and intellectual control-rather than debating whether AI was used at all.
This approach preserves flexibility as technology evolves. It also keeps copyright tied to genuine human creativity rather than weakening the concept by protecting machine-generated output the same way.
A special category for pure machine output
Bangladesh should avoid granting full copyright protection to purely autonomous AI-generated works where meaningful human creativity is absent. Doing so would gradually erode the philosophical foundations of copyright itself.
Instead, the country might eventually establish a limited legal category-known as sui generis protection-for certain autonomous AI outputs. This would recognise the value of machine-generated content without treating it as equivalent to human intellectual creation.
Copyright law now must confront an uncomfortable question: if expressive works can be generated without meaningful human creativity, what exactly is copyright protecting anymore? Bangladesh has an opportunity to answer that question with a balanced framework that encourages innovation while preserving the human foundations copyright has always depended on.
Legal professionals working on copyright and intellectual property issues may find it useful to understand the technology at the centre of this debate. Resources on Generative AI and LLM and AI for Legal can help build that foundation.
Your membership also unlocks:
