LAGOS/Nigeria: Nigeria risks becoming vulnerable to decisions made by foreign artificial intelligence companies unless it develops stronger local capacity to evaluate and regulate advanced AI systems, Orodata Science has warned.
The Nigerian data science and research organisation made the observation in a policy brief titled “Pacing Frontier AI: Protecting Rights and Public-Interest Access in Nigeria,” released in Lagos on Wednesday, urging civil society organisations (CSOs) to become more actively involved in the emerging debate over how increasingly powerful AI systems should be governed.
Orodata said Nigeria’s growing reliance on foundational AI models developed and controlled outside Africa could expose local businesses, public institutions and civic initiatives to risks beyond their control.
It noted that changes in the pricing, licensing arrangements or availability of foreign AI platforms could have direct consequences for Nigerian start-ups and public-interest projects, including initiatives in healthcare, education and governance.
The organisation therefore called for greater attention to data sovereignty, independent AI evaluation and the development of Nigerian expertise capable of testing advanced AI systems against local realities.
Orodata’s intervention comes amid an expanding international debate over whether the development of frontier AI should be deliberately paced to allow safety mechanisms and regulatory systems to catch up with rapidly advancing capabilities. The debate gained renewed attention following proposals by Anthropic CEO Dario Amodei for stronger independent evaluation, technical safety requirements and international coordination.
According to Orodata, Nigeria should not approach the debate merely as a question of whether AI development should be slowed or accelerated. Instead, it argued that policymakers should establish safeguards that allow the country to benefit from AI while limiting risks to citizens, institutions and democratic processes.
The organisation said “pacing” should not be interpreted as an outright suspension of software development or the withdrawal of existing digital services.
Rather, it involves introducing regulatory checkpoints, independent red-teaming and capability thresholds before increasingly advanced AI systems are trained or deployed.
Orodata identified independent assessment, harmonised safety standards and international cooperation as important components of responsible frontier AI governance.
It argued that independent evaluators must have sufficient access to AI development processes to assess systems throughout their training and deployment lifecycle.
The organisation also warned that voluntary commitments by technology companies should not be regarded as substitutes for enforceable regulation.
It urged Nigerian civil society groups to use existing instruments, including the Nigeria Data Protection Act 2023, the National Artificial Intelligence Strategy and the African Union Continental AI Strategy, to demand measurable commitments from AI developers.
Among the safeguards proposed were support for African-led red-teaming exercises, stronger protection of locally generated data and more equitable access to AI technologies.
Beyond the risks associated with highly advanced systems, Orodata said policymakers should also address the everyday harms already associated with AI deployment.
It identified political disinformation, deepfakes, financial fraud, linguistic and cultural exclusion and the exploitation of African data workers among issues requiring greater regulatory attention.
The organisation said advanced AI systems should be tested for their performance in Nigerian languages and local socio-political environments, as well as their ability to handle the country’s ethnic diversity and locally specific fraud patterns.
It consequently called for increased investment in Nigerian universities, research institutions and civil society organisations to enable them to undertake AI safety research and red teaming based on African circumstances.
Orodata further urged international donors, development finance institutions and technology companies to provide dedicated financial and technical support for such work.
It said Nigerian CSOs should also be given greater access to international AI safety networks and consultations so that perspectives from the Global South would be represented in decisions shaping the technology’s future.
On regulation, the organisation advocated a tiered liability system that distinguishes between those developing frontier AI systems and organisations deploying them.
It said developers should be responsible for structural design problems, baseline safety and inadequate testing, while organisations deploying AI should bear obligations proportionate to the potential consequences of their applications.
Orodata also called for meaningful human oversight whenever automated systems are used in decisions with significant consequences for individuals or communities.
It cited areas such as social welfare allocation and legal or community interventions as examples where human involvement should not be eliminated.
For non-profit organisations already using AI tools, the organisation recommended five immediate safeguards: auditing AI applications, preventing sensitive beneficiary information from being entered into consumer AI systems, requiring human verification of AI-generated outputs, securing clear contractual terms on data retention and security, and participating in civil society knowledge-sharing networks.
The organisation maintained that Nigeria did not need to wait for a global agreement or the passage of additional domestic legislation before taking practical steps to protect citizens and institutions from AI-related risks.
It said the objective should be to develop an AI governance framework that protects democratic rights and social justice without shutting African countries out of the benefits of technological advancement.
Orodata urged Nigerian CSOs to use the current debate as an opportunity to press for stronger data protection, greater domestic technical capacity and affordable access to AI for public-interest purposes.
It stressed that voluntary restraint by executives of foreign technology companies, on its own, could neither guarantee Nigeria’s digital sovereignty nor adequately protect civic space from algorithmic harms.
