AI vs. Authors: The $1.5 Billion Anthropic Copyright Showdown and What It Means for Tanzania
The Anthropic Copyright Case
The Anthropic settlement, in the landmark copyright lawsuit Authors Guild v. Anthropic, resulted in a $1.5 billion payout. This case centered on the allegation that the AI startup Anthropic had used a massive dataset of pirated books—downloaded from unauthorized sources like Library Genesis and Pirate Library Mirror—to train its large language models. The authors argued that this practice constituted copyright infringement by reproducing their works without permission. The settlement required Anthropic to compensate the affected authors and to destroy the infringing datasets. This case sets a significant precedent for the AI industry, underscoring the legal risks of using unauthorized, copyrighted material as training data.
Comparison to Tanzanian Law
Tanzania’s copyright framework is governed by the Copyright and Neighbouring Rights Act, Cap. 218 R.E 2025. This Act provides legal protection to authors who are natural persons (humans) for their original literary and artistic works. The law grants authors exclusive rights, including the right to reproduce, publish, and distribute their work.
A key difference between the Anthropic case and the current state of Tanzanian law is the lack of specific legislation addressing AI. The Tanzanian Act, with its focus on human authorship, has not yet been updated to address the complexities of AI-generated content or the use of copyrighted data for training AI models. This creates a legal gray area where it is unclear how existing copyright principles would apply to the activities of an AI company.
Implications for AI Models Trained in Tanzania
The Anthropic case holds significant implications for AI models trained in Tanzania, highlighting key legal risks and challenges.
- Legal Uncertainty: Due to the absence of specific AI legislation, companies training AI models in Tanzania face legal ambiguity. While the law protects a human author’s right to reproduce their work, it does not explicitly state whether the use of such works for AI training constitutes a violation.
- Risk of Infringement: Despite the legal ambiguity, an argument could be made under the current law that training an AI model on copyrighted material infringes on the exclusive rights of the author. A Tanzanian court could, in theory, find that the act of copying and using a work for training purposes violates the author’s right to reproduction.
- No Copyright for AI-Generated Works: The Tanzanian Act defines an author as a “natural person,” meaning any work created autonomously by an AI model would not be eligible for copyright protection. This could disincentivize the development of creative AI models within the country, as developers and users would not have legal ownership over the works produced.
Implications for Copyright Law in Tanzania
The Anthropic case and global trends in AI technology underscore a pressing need for legal reform in Tanzania’s copyright framework.
- Pressure for Reform: The global spotlight on AI and copyright, spurred by cases like Anthropic’s, places pressure on countries to modernize their laws. Legal experts and stakeholders in Tanzania have already been advocating for a comprehensive AI policy and amendments to the Copyright and Neighbouring Rights Act RE 2025 to address this legal gap.
- Harmonization with International Standards: As a member of the World Intellectual Property Organization (WIPO) and a signatory to the Berne Convention, Tanzania is obligated to provide robust copyright protection. Updating its laws to account for AI would align the country with international legal developments and ensure its framework remains relevant in the digital age.
- Future Legislation: It is highly likely that Tanzania will need to enact new legislation or amend existing laws to clarify the legal status of AI-generated content, establish clear rules for the use of copyrighted data in AI training, and determine the rights and liabilities of AI developers and users. Without such reforms, the current framework may stifle innovation and fail to adequately protect the rights of creators.
Author and Sources
This review is based on an analysis of the provided Financial Times article and on Tanzanian copyright law.
- Author: Sunday Ndamugoba, Partner at RIVE & Co.
- Contact: sunday@rive.co.tz
Sources:
- Financial Times article
- Tanzania’s Copyright and Neighbouring Rights Act, Cap. 218 R.E 2002
- Various articles and legal analyses of Tanzanian intellectual property law
Disclaimer
- Disclaimer: This is a general review for informational purposes only and does not constitute legal advice. For specific legal guidance regarding intellectual property and AI in Tanzania, please consult with a qualified legal professional.