The Mirror Across the Atlantic
On June 24, 2026, a coalition of 35 US newspaper publishers operating nearly 400 regional and local outlets across 33 states filed suit in the Southern District of New York against OpenAI and Microsoft. The complaint alleges that the companies systematically crawled publisher websites — including paywalled content — copied articles to their own servers, stripped out copyright management information, and fed the resulting datasets into ChatGPT and Microsoft Copilot without permission or payment. The case is the largest coordinated legal action by local and regional news organisations against an AI developer to date.
Whether or not the publishers prevail, their lawsuit is a landmark act of legal self-defence that most of the world's content creators cannot replicate. Nigerian newspaper editors, Nollywood producers, and Afrobeats artists face the same economic threat — AI companies indexing and training on their work — but possess no comparable legal instrument to wield.
What the US Publishers Are Claiming
The complaint rests on two pillars. The first is direct copyright infringement: OpenAI and Microsoft made unauthorised copies of protected works on a massive scale. The second is a violation of the Digital Millennium Copyright Act, which prohibits the removal of copyright management information — author bylines, publication names, copyright notices — from protected works. The publishers seek statutory damages, injunctive relief, and potentially the deletion of infringing model weights.
The case joins a crowded docket. The New York Times filed against OpenAI and Microsoft in December 2023 and is in active discovery as of mid-2026. The common thread across all these suits is the same question: does training an AI system on copyrighted material constitute infringement, or does it qualify as fair transformative use? That question is genuinely contested. The strongest argument for AI companies is that training produces a statistical model, not a copy, and that restricting training data access would chill AI development. This is a serious argument: overly broad copyright protection could entrench incumbents who have already scraped the web and freeze out new entrants. Courts have not yet resolved the matter, and a ruling either way will reshape the economics of AI development globally.
Nigeria's Legal Gap
Whatever the outcome in New York, Nigerian creators cannot bring a comparable case under domestic law. Nigeria's Copyright Act 2022 (Act No. 8 of 2022, commenced March 17, 2023 and registered by WIPO as legislation NG052) is the country's most significant copyright modernisation in decades. It updated digital rights, addressed piracy enforcement, and expanded neighbouring rights. What it did not do is address AI.
The Act contains no explicit provision governing text and data mining (TDM) — the process by which AI systems extract patterns from large corpora of copyrighted material. The closest provision is Section 20(1)(o), which creates a fair dealing exception for "transient and incidental reproductions" integral to a "technological process." Legal analysts have concluded that this exception — designed for caching and routing in telecommunications networks — cannot plausibly extend to large-scale AI training ingestion, which is neither transient nor incidental. The recommendation from intellectual property practitioners is unambiguous: seek rightholder permission before using copyrighted works for AI training, because no statutory safe harbour exists.
The Nigerian Copyright Commission, the statutory body responsible for administering the Act, has published no guidelines, advisory opinions, or policy statements specifically addressing AI training. In the absence of regulatory guidance, any Nigerian publisher or artist whose work is scraped and ingested into a commercial AI model has no clear domestic pathway to seek compensation or injunctive relief.
Who Is Exposed
The practical exposure is substantial across three sectors.
Nigerian news publishers produce original investigative and policy journalism that circulates widely online and is indexed by AI training pipelines. Unlike their US counterparts, they have no DMCA equivalent to invoke, no published Commission guidance to cite, and no precedent case to build on.
Nollywood, Africa's largest film industry by output, generates scripts, dialogue, and audiovisual content that AI systems can draw upon for training synthetic media generation. Industry legal commentary has flagged that the Copyright Act's protections were designed before AI disruption and are now inadequate for audiovisual works — particularly as AI models capable of generating Nollywood-style scripts and cloning Nigerian voices become commercially available.
Afrobeats artists face an acute version of this threat at the voice and composition layer. In 2025, an AI-generated simulation of Burna Boy circulated on social media; AI detection services confirmed the audio was entirely synthetic. Voice cloning models are trained on recordings — recordings protected by copyright under the Act — yet no Nigerian enforcement mechanism currently addresses this specific form of infringement.
None of these creators currently has a mechanism to opt out of AI training crawls, to demand compensation if their work is used, or to bring a domestic copyright claim analogous to the US publishers' suit.
The International Baseline Nigeria Is Missing
Other jurisdictions have handled this explicitly. The EU's Copyright in the Digital Single Market Directive (2019) introduced a text-and-data-mining exception for research purposes alongside a commercial TDM reservation that lets rightholders opt out. The UK, Singapore, and Japan have each enacted their own TDM frameworks, differing in scope but sharing the principle that the legal relationship between AI training and copyright does not resolve itself through general fair use doctrine alone — it requires explicit legislative choices.
Nigeria has neither a TDM carve-out nor a defined opt-out mechanism. The National Digital Economy and E-Governance Bill targeting passage in 2026 focuses on AI standards, data governance, and regulatory sandboxes. In its current draft, it does not address copyright-specific AI issues. That is a gap the National Assembly should close before the bill passes.
A Proportionate Path Forward
The right reform is not to ban AI training on Nigerian content — that would be unenforceable, would disadvantage Nigerian AI researchers who need training data, and would push AI companies to simply route their pipelines through jurisdictions with no restrictions. The right reform is a TDM framework modelled on the EU's: a research exception paired with a commercial opt-out right that gives Nigerian publishers, Nollywood producers, and music rightholders a legal instrument they can actually use.
The Nigerian Copyright Commission should follow this with guidance clarifying whether AI training on copyrighted content without a TDM licence constitutes infringement under the Act — or, if it concludes otherwise, explain the reasoning. Silence is the worst possible answer for creators watching foreign AI companies index their work in real time and facing no domestic recourse.
The US publishers filing in New York are doing what any property-rights holder would do when their assets are taken without compensation. Nigerian creators deserve the same option. Right now, the law does not give it to them.