Recently, some Nigerian male users on X (formerly Twitter) began using Grok, an artificial intelligence (AI) tool developed by Elon Musk, to digitally harass women by asking it to undress them. This is not unique to Nigeria. Earlier this year, deepfake explicit images of popular singer Taylor Swift went viral.

Beyond harassment, AI has been misused globally to amplify biases, worsen inequalities, spread misinformation, and institutionalise injustice. For instance, facial recognition tools in the United States (U.S.) have wrongly identified Black people, leading to false arrests. AI has also been weaponised to spread election misinformation in both Nigeria and the U.S.

These incidents highlight how AI can be abused worldwide and underscore the urgent need for responsible regulation and governance frameworks. Countries such as the European Union, Singapore, and Canada are already moving in this direction.

The EU’s AI Act categorises AI uses into “unacceptable,” “high risk,” “limited risk,” and “minimal risk.” In contrast, the U.S. framework relies on litigation, empowering citizens to sue corporations for AI misuse.

Nigeria, however, cannot simply import these models. Unlike the EU, Nigeria struggles with weak institutional enforcement. And unlike the U.S., Nigeria lacks a robust legal system that empowers citizens to hold powerful firms accountable.

Currently, Nigeria has no dedicated law regulating AI. In 2022, the National Information Technology Development Agency (NITDA) launched stakeholder consultations on a National Artificial Intelligence Policy (NAIP), with a first draft completed by March 2023. By August 2024, the Minister of Communications, Innovation, and Digital Economy released the draft National Artificial Intelligence Strategy (NAIS), outlining Nigeria’s roadmap for ethical and responsible AI adoption.

As Nigeria builds a regulatory framework, it must avoid copying foreign models wholesale. Local realities such as ethnic diversity, digital literacy levels, and the enforcement capacity of governance institutions must shape Nigeria’s approach.

A responsible Nigerian AI framework should rest on five pillars: risk assessment, bias mitigation, transparency and audit, redress mechanisms, and regulatory compliance.

High-risk AI systems capable of harming Nigerians or threatening national stability should face strict oversight before deployment, while low-risk systems can be encouraged under lighter regulations. AI must also be tested against Nigeria’s diverse social realities to mitigate bias.

Since most AI systems are trained on Western data, Nigeria must ensure that deployed systems are trained on local datasets reflecting ethnic, gender, linguistic, and socioeconomic diversity. Organisations using AI should be required to publish audit reports detailing risks and safeguards.

Passing an AI law is not enough. Strict enforcement is critical. Nigeria has a history of policies that remain unenforced, reduced to promises on paper. A toothless AI law would be no different. Citizens must also have clear avenues to challenge harmful AI outcomes, such as being unfairly denied a loan, misclassified by facial recognition, or unjustly targeted by flawed policing systems.

This calls for a comprehensive strategy: investing in indigenous digital infrastructure, training AI models on Nigerian datasets, and boosting digital literacy across the population.

The stakes are high. Adopting foreign regulatory frameworks without adaptation or failing to adopt any at all could have dire consequences. Once harmful AI systems are embedded, reversing them is difficult and risks eroding public trust in technology.

AI can either empower or endanger Nigeria’s future. By crafting frameworks that fit local realities and learning from global mistakes, Nigeria can build public trust and ensure that AI works for Nigerians, not against them.