Intellectual Property in the Age of AI: Challenges for Bangladesh

Introduction:

The rapid advancement of artificial intelligence has fundamentally disrupted traditional intellectual property modelsacross the world, putting a historically unprecedented burden on human-centered legal frameworks based on creativity and innovation. Bangladesh is currently undergoing accelerated economic transformation and incremental IP law reform, particularly which make these challenges are too difficult to manage. The emergence of AI technologies raises new questions about fundamental IP principles like authorship, inventorship, originality, and ownership that the existing legal paradigms are not adapted to address. Such concerns are additionally complicated in Bangladesh by the availability of factors like limited infrastructure for technology, rural-urban digital divides, embryonic stage take-up of AI in different sectors, and balancing innovation incentives with access to technology in the context of a developing economy. This blog examines how rapidly advancing AI technologies are challenging Bangladesh’s traditional intellectual property laws, especially questions of authorship, ownership, and patentability. It highlights the legal gaps, enforcement limitations, and policy reforms needed as the country prepares for full Trade-Related Aspects of Intellectual Property Rights (TRIPS) compliance after Least Developed Country (LDC).

Evolution of IP Laws in Bangladesh:

The intellectual property regime of Bangladesh has undergone a significant change in the recent years, moving from colonial law to modern, harmonised international law. The Patents and Designs Act 1911 was replaced by the Patent Act 2023, operational since February 27, 2025, with an extended patent term from 16 to 20-years aligned with global standards and utility model patents for incremental innovation. Similarly, the Copyright Act 2023, replaced the Copyright Act 2000 with stricter penalties for infringement and enhancing protection for digital works.

Global Perspectives on AI and IP:

Around the world, governments have taken various approaches to tackling AI-related IP concerns. The European Union's AI Act 2023 indirectly supports the management of AI-generated works and copyright attribution. Germany granted limited rights to AI systems, acknowledging their role as creators. With the exception of most patent offices, including those that handled the DABUS patent applications, all have declined to acknowledge AI systems as inventors, citing patent laws demanding human inventorship.

The doctrine of  traditional fair use must be understood in the new environment of AI-generated material and data-intensive analysis, with questions of what precisely constitutes "transformative use" where there are AI-enabled tools involved. These global developments leave Bangladesh facing the challenge of designing an IP system capable of supporting AI-driven innovation without leaving creators vulnerable.

Current Legal Framework:

i) Patent Law and AI Inventions

The Bangladesh Patent Act 2023 retains the human inventorship requirement, and the law explains "there is no scope for AI to be appointed as inventor under the said Act". This creates challenges in AI-generated inventions where human contribution may be limited or difficult to define. Although, the Act incorporates TRIPS flexibilities to reduce challenges such as evergreening and drug price extravagance, but lacks clear provisions for evaluating the patentability of inventions concerning AI. Determining the “inventive step” becomes more complex to ascertain when AI systems possess enormous knowledge and capabilities that can exceed human know-how in some fields. The problem of determining "obvious to a person skilled in the art" may requires reconsideration, as AI systems can analyse massive prior art datasets and generate novel combinations more efficiently than human inventors.

ii) Copyright Law and AI-Generated Content

The Copyright Act 2023 defines "author" as the creator of the work but does not address authorship for AI-generated content. This leaves the issue of authorship and ownership of works created by AI systems in limbo, particularly when AI systems create content with minimal human involvement. The Act protects literary works, computer works derived from information technology, and artistic works under section 14(1)(a), 14(1)(b), and 14(1)(c) respectively, yet it does not provide a specific framework for works produced with the assistance of AI.

iii) Enforcement Mechanisms and Challenges

Enforcement of IP in Bangladesh is carried out through civil remedies, criminal sanctions, and administrative measures. The Department of Patents, Designs and Trademarks (DPDT) under the Ministry of Industries and Copyright Office under the Ministry of Cultural Affairs oversee IP regulation. However, enforcement faces gigantic challenges like limited resources, a lack of specialised IP courts, and insufficient digital rights management mechanisms.

The system of copyright enforcement provides six months to four years' imprisonment and fines for infringement, with harsher penalties for software piracy. However, these measures were designed for traditional infringement and may be inadequate for addressing AI-specific issues such as unauthorized use of training data or algorithmic replication.

Principal Challenges in the Age of AI:

i) Authorship and Ownership Determination

One of the key challenges facing Bangladesh's IP system pertains to determining authorship and ownership of works produced by AI. Current law requires human creators for copyright and human inventors for patent rights, yet AI systems are generating increasing amounts of content and inventions with minimal or no human touch. This creates legal uncertainty that can discourage investment in AI Research and Development (R&D) or lead to IP ownership conflicts.

The challenge is particularly stark for productions made by large language models or creative AI programs that produce literature, art, music, or other works traditionally copyrightable. In the absence of authoritative clarification from the law, creators, companies, and investors remain uncertain about their rights over AI-generated content.

ii) Training Data and Copyright Issues

AI systems generally have to be trained on massive amounts of data, often containing copyrighted content without explicit permission. AI systems that use copyrighted content for training are not expressly covered by Bangladesh's Copyright Act, 2023. This creates a potential risk of liability for organisations engaged in the development of AI systems utilizing copyrighted content within their training data sets.

iii) Digital Divide and Access Issues

Bangladesh also faces significant digital infrastructure problems that affect AI development and intellectual property protection. Penetration of the internet has a huge rural-urban gap, with only 36.5% of rural residents using the internet compared to 71.4% in urban communities. This kind of digital gap limits exposure to AI technologies and leads to innovation access imbalances across different regions. The thin digital environment also has a bearing on enforcing IP, since monitoring and detection of infringement become increasingly difficult in low-connectivity areas.

iv) International Compliance and TRIPS Obligations

Bangladesh's graduation from LDC status within 2026 will entail strict enforcement of TRIPS agreement obligations, ending current exemptions for pharmaceutical patent and other IP safeguards. It is a period of speedy AI development, and the country needs to strike a balance between innovation incentives and access to technology concerns.

The graduation also affects Bangladesh's ability to maintain certain flexibilities in IP law that have helped domestic industries, with the pharmaceutical industry being the most notable. The nation must balance maintaining continued access to essential technology with offering adequate intellectual property protection for AI innovations, given the expanding use of AI in healthcare and other fields.

Policy and Regulatory Responses:

i) National AI Policy 2024

Bangladesh's National AI Policy 2024 provides guidance on AI governance, adoption, and development in government ministries. The policy is centered around the education, healthcare, transport, and agriculture sectors, but provides very limited guidance on IP-related matters. Though the policy mentions the establishment of a National Artificial Intelligence Center of Excellence (NAICE) and monitoring committee to ensure that ethics are dealt with, the policy lacks certain provisions to deal with IP matters during AI development.

The policy does not clarify approaches to handling AI-generated material, protecting creators' rights, or dealing with IP issues in AI implementations. The failure leaves fundamental questions unresolved as Bangladesh continues its AI adoption initiatives.

ii) Cyber Security and Data Protection Framework

The Cyber Security Ordinance 2025 and amendments to data protection laws attempt to bring privacy and computer crime legislation up to date but have little to say about AI-generated material or algorithmic decision-making. The framework does include breach reporting requirements and cross-border transfer restrictions, but algorithmic transparency, liability, and human oversight are not addressed. These regulatory loopholes leave AI users and developers in doubt about their obligations when utilising AI that can generate IP-protected content or handle copyrighted material.

iii) Industrial Property Administration System (IPAS)

The government's launch of the Industrial Property Administration System (IPAS) 4.0 software in May 2025 is a step towards the modernisation of IP administration. Through automation, the system being created with assistance from the World Intellectual Property Organisation (WIPO) seeks to elevate Bangladesh's patent, design, and trademark services to an international standard. Nevertheless, the system was not expressly created to handle IP applications linked to AI or the particular problems they pose.

Comparative Analysis and International Best Practices:

i) Singapore's Approach

Singapore's Intellectual Property Office (IPOS) released comprehensive guidelines for AI-related patent filings in 2024, looking to eligible subject matter for patents and requesting "actual contribution" tests in the event of AI inventions. This approach tries to test what the inventors have "really contributed to human knowledge" rather than focusing on the application of AI tools.

ii) European Union Framework

The EU framework emphasises openness, responsibility, and human oversight within AI systems, trajectories that can inform Bangladesh's policy for AI-related matters of IP.

iii) Lessons for Bangladesh

Global best practices set the requirement for Bangladesh to enact specific guidelines on AI-related IP issues, establish specialised courts or procedures for AI cases, and enact clear guidelines for evaluating AI contribution to inventions and creative work. The country would be served well by a balanced approach of protecting creators' rights but allowing AI innovation.

Economic and Innovation Implications:

i) Impact on Innovation Ecosystem

Bangladesh ranks 106th out of 133 economies in the Global Innovation Index 2024, with particular weaknesses in human capital and research (128th) and business sophistication (126th). Uncertainty of IP rights on AI may further deter the development of innovation since uncertain legal frameworks discourage investment in AI R&D.

The country's innovation outcomes (92nd globally) are more robust than innovation inputs (114th), showing potential for innovative output which could be enhanced with stronger differentiated IP protection for such works generated by AI. However, without appropriate legal means, this potential may remain unrealised.

ii) Technology Transfer Challenges

Bangladesh is constrained by IP asset commercialisation and technology transfer, particularly in AI technology. Lack of proper university-industry linkages and lack of full-time researchers engaged in R&D are hindrances to effective innovation systems. These are further exacerbated by ambiguous intellectual property rights over innovations produced by AI, which may deter international investment and knowledge transfer.

iii) Pharmaceutical and Healthcare AI

As Bangladesh's pharma industry prepares to face TRIPS-readiness after LDC graduation, AI technologies in pharmaceutical innovation, diagnostic tools, and precision health present both opportunity and risk. The industry's strong strengths in generic drug manufacturing would potentially be transferable to AI-powered pharmaceutical innovation, but ambiguous IP regimes will limit this potential.

Conclusion:

Bangladesh urgently needs an intellectual property framework equipped to meet the demands of AI-driven innovation. Although, the recent legal reforms through the Copyright Act 2023 and Patent Act 2023 are important milestones but fall short in addressing AI-specific issues such as authorship, ownership, training data, enforcement mechanisms, and digital infrastructure limitations. The challenges are compounded by Bangladesh's impending graduation from the LDC status in 2026, when there would be the imperative of full TRIPS compliance and maintaining innovation incentives and access to technologies. Furthermore, the digital divide and limited judicial capacity to address complex AI-related disputes add additional layers of difficulty. The path forward entails not only legal and regulatory reform but also investment in technological infrastructure, building human capacity, and global collaboration.