When Machines Create, Who Owns the Rights? AI and the Legal Vacuum in Bangladesh
Mst Farjana Khatun
Published: 19 Nov 2026, 01:08 AM
Not too long ago, the idea of a machine writing poetry,
composing music, painting, or building new technology belonged strictly to
science fiction. Today, generative artificial intelligence has made all of that
part of our everyday lives. Tools like ChatGPT, Gemini, Midjourney, and Sora
can produce complex writing, artwork, and technical designs within seconds.
Often, they do this without any direct human involvement.
While this incredible advancement in technology has led to
unimaginable speed and economic opportunities, it has caused a serious legal
quandary across the globe. This is particularly true in the area of
Intellectual Property (IP). At the root of this crisis lies a seemingly simple
but very complex question: When a machine creates something original, who will
own it?
International intellectual property law is based on a
strict "human-centric" approach. Key global treaties like the Berne
Convention (1886), TRIPS Agreement (1995), and WIPO Copyright Treaty (1996)
state that an "author" or "inventor" must be a human. The
system exists to reward human intellect, effort, and creativity with legal
protection. Because these treaties were developed long before modern algorithms
and AI, they offer no clear guidance on "non-human authorship". As a
result, countries are trying to interpret their own domestic laws, leading to
legal conflicts globally.
Recently, courts and IP offices across several
jurisdictions have issued key rulings that highlight the clash between AI
capabilities and traditional laws. In the DABUS patent cases, American computer
scientist Dr. Stephen Thaler used his AI system, “DABUS,” to design a
specialized food container and an emergency warning light without human help.
He filed patent applications in more than a dozen countries, listing “DABUS"
as the sole inventor. In 2023, the UK Supreme Court dismissed his appeal,
ruling that under the Patents Act 1977, an inventor must be a human being. Like
the UK Supreme Court, the USA and the EU also rejected his applications.
Similarly, in Thaler v. Perlmutter (US, 2023), a US federal
court upheld the rejection of copyright registration for an artwork created
entirely by AI, noting that human creativity is essential for copyright
protection. Without direct human creative control, AI-generated works fall into
the public domain. Meanwhile, in The New York Times v. OpenAI and Microsoft
(US, 2023), The New York Times sued OpenAI and Microsoft, alleging they used
millions of its copyrighted articles to train models like ChatGPT without
permission or payment. While the companies claim fair use, creators view it as
large-scale intellectual property theft.
Bangladesh is rapidly growing its digital economy, IT
sector, and creative industries. Also, AI-related legal issues are becoming
increasingly relevant here. Recently, the government introduced key laws and
policies such as the Copyright Act, 2023, the Bangladesh Patent Act, 2023, the
Personal Data Protection Act, 2026, and the Draft National AI Policy 2026–2030.
Even with these steps, a clear legal vacuum remains where AI and intellectual
property meet.
First, there is ambiguity over “author” and “originality”.
While the Copyright Act, 2023 protects literary, artistic, and computer-based
works, it offers no framework for generative AI. Because the law still requires
traditional human authorship, it remains unclear how much human prompting is
needed to claim copyright over an AI-assisted work.
Second, under the Bangladesh Patent Act, 2023, an invention
must come from a human "inventor" to qualify for a patent. Since the
new rules under this Act are still unfinalized and the country relies on the
older 1933 Rules, uncertainty remains over whether AI-assisted software
innovations can be patented. Meanwhile, the Draft National AI Policy
(2026–2030) is not yet a binding law, and existing copyright laws offer no
clear remedy when AI models train on the works of Bangladeshi writers,
journalists, or artists without permission.
Bangladesh cannot let its laws lag behind technology, nor
should it block innovation with strict rules. Through legal updates, the
government should adopt a “Human-in-the-Loop” rule, meaning AI-assisted works
with real human input get IP protection, while purely machine-made ones do not.
Tech companies must also disclose their training data and pay fair royalties to
local creators, alongside training judges and IP officials to handle AI
disputes.
Artificial intelligence is already reshaping how we write, design, code, and innovate. Updating Bangladesh’s intellectual property laws is not about giving rights to algorithms. Rather, it is about protecting human creativity and helping our creative and tech sectors grow safely in the age of AI.
The write is a student of Law at European University of
Bangladesh