Nigeria won a prize for responsible AI, then went after the AI companies

WHAT HAPPENED

On 6 July, President Tinubu directed the Federal Competition and Consumer Protection Commission to investigate X, Meta, Alphabet and a set of generative AI platforms. The trigger was a joint petition from the Nigerian Press Organisation, the umbrella body for the country's newspaper proprietors, journalists, broadcasters and online publishers. The accusation: that these companies scrape copyrighted Nigerian journalism to train models without permission or payment, dominate digital advertising, and refuse to negotiate commercial terms with local publishers.

The FCCPC's Tunji Bello promised an independent, evidence-based inquiry and was careful to say it presumes no wrongdoing. The commission has form. It fined Meta $220 million in 2024 over consumer protection and data privacy. Meta is appealing.

Three days later, the Global Center on AI Governance published the second Global Index on Responsible AI. Nigeria came 38th of 135 countries with a score of 45.93, first in Africa, ahead of Egypt. In 2024 it ranked 80th on 7.21. It was also named a global bright spot for AI literacy.

WHY THIS MATTERS

Read together, these are not a contradiction. They are the same policy. A country that intends to govern AI rather than merely host it has to do the unglamorous half: send the competition regulator after the largest companies on earth on behalf of its newspapers.

Nigeria's index score was driven by exactly this pairing. Most countries chase adoption and skills. Nigeria paired those with legal safeguards. The probe is what a safeguard looks like when it stops being a PDF.

For anyone who makes a living from words in Kenya or anywhere else on the continent, this is the live question. Your work is training data. Nobody asked you and nobody paid you. Nigeria is the first African government to formally ask whether that is legal.

THE PART NOBODY IS CELEBRATING

Now look at the number again. 45.93 out of 100. That is a failing mark in any classroom in Lagos or Nairobi, and it is the best on the continent. The global average is roughly 35. The index found that evidence of actual implementation exists in only about 55% of countries with responsible AI frameworks, and just 45% across the Global South.

So the honest reading of Nigeria's win is not that Nigeria has solved AI governance. It is that almost nobody has, and Nigeria wrote more of it down. A strategy document scores. Enforcement is harder, and the FCCPC is about to find out how hard, because Meta is still fighting the last fine two years on.

The ranking is the promise. The probe is the test. Watch the probe.