Highlights
- AI is not theft by default as studying Tanzanian works is observation, not infringement under the Copyright Act.
- Fair use is the real battleground as COSOTA Regulations demand proof of market harm before AI use can be restricted.
- Creators must guard their style as the risk lies in AI outputs that mimic and not the input prompt.
1. Introduction:
In Dar es Salaam’s bustling creative hubs, Bongo Flava artists, Swahili novelists, and Silicon Zanzibar’s tech entrepreneurs share a common anxiety, is artificial intelligence stealing their creativity? Musicians worry that their rhythms will be replicated by algorithms, writers fear their unique style will be imitated, and startups wonder if their innovations will be absorbed into global AI systems without recognition. These concerns are not unfounded, but they also reveal a misunderstanding of what copyright law in Tanzania is designed to do.
Copyright is not meant to block creativity; it exists to channel it, to balance protection with access, and to ensure that knowledge can be shared while creators are rewarded. When AI systems study publicly available material, they are engaging in observation and analysis. These acts do not amount to infringement. The real legal questions arise when AI systems store material for longer periods or generate new works that compete with originals. This is where Intellectual Property Law Tanzania, particularly the Copyright and Neighbouring Rights Act (RE 2023), provides guidance.
2. Legal Framework and Evidence
a. The Copyright and Neighbouring Rights Act (RE 2023)
Tanzania’s copyright law protects authors, musicians, and artists, but it also allows exceptions for education, research, and fair use. Section 12 of the Act permits reproduction for teaching and research, which can extend to AI analysis. This means that an AI system studying Bongo Flava rhythms or Swahili literature is engaging in observation and not infringement. The law already anticipates that knowledge must be studied to grow, and AI can be seen as an extension of this principle.
b. International Case Law Applied Locally
International precedents help clarify how Tanzanian courts might approach AI. In Cartoon Network LP v. CSC Holdings (Cablevision), the U.S. court held that storing data for only 1.2 seconds was too short to be considered a “copy.” Applied in Tanzania, temporary AI storage should not be treated as infringement under the Copyright and Neighbouring Rights Act.
Similarly, in Authors Guild v. Google Books, scanning entire books to create a searchable database was deemed fair use because the use was transformative. This reasoning applies directly to Tanzania, where AI could analyze Swahili texts or Tanzanian case law to improve language tools or legal research databases. The courts emphasized that fair use depends on whether the new use harms the market for the original. For Tanzanian authors and musicians, the real issue is not whether AI studies their works, but whether AI-generated products compete unfairly with their originals.
c. COSOTA Regulations and Local Context
The COSOTA Regulations reinforce this balance between protection and access. Just as students may study copyrighted works for learning, AI systems can analyze them without reducing their market value. The real challenge arises when AI generates works that mimic Tanzanian creators. In such cases, the law provides remedies to challenge unfair competition.
3. Strategic Advisory
Policymakers should focus on clarifying how the Copyright and Neighbouring Rights Act (RE 2023) applies to AI rather than rushing to draft new legislation. The existing fair use provisions can be extended to AI, ensuring both innovation and protection. Judges must be trained to apply the four fair use factors rigorously: is the AI use transformative, what is the nature of the work, how much of the original is used, and does the AI product harm the market for the original?
Creators should monitor AI-generated outputs rather than the learning process. When AI studies their works, it does not reduce their value. The real issue arises if AI produces works that closely imitate their style or substance. In such cases, creators can rely on existing remedies under Tanzanian law. Registering works with COSOTA, documenting originality through invoices and contracts, and cultivating a clear Tanzanian market presence are all practical steps to strengthen protection.
Innovators should design AI systems that transform knowledge rather than replicate it. Systems that generate new insights. For example, analyzing Swahili proverbs to build better translation tools, would be legally safer and socially beneficial. Engaging with regulators early and demonstrating compliance with Intellectual Property Law Tanzania will build trust and reduce risk.
4. Conclusion
AI is not inherently harmful to Tanzanian creativity. Observation and analysis are not infringement and so generated works must be judged case by case. By grounding international precedents in Tanzanian law and applying the COSOTA Regulations, Tanzania can ensure that AI strengthens rather than undermines the sharing of knowledge and creativity. The law already provides the tools, what is needed now is strategic application and awareness.
Mouline Ruhangisa is a lawyer at RIVE&Co in Dar es Salaam. She is reachable at ip@rive.co.tz or moulin.rive@outlook.com
Disclaimer
The contents of this publication are intended for general information purposes only and do not constitute legal, tax, or professional investment advice. We strongly recommend that parties seek bespoke legal counsel to navigate the complexities of the Tanzanian laws.