The AI Race Is Now a Political Contest and Africa Risks Becoming a Rule-Taker

Governments are moving to regulate artificial intelligence as concerns grow over jobs, copyright, misinformation, data control and national security. Africa’s challenge is to protect its interests without closing the door on innovation. Artificial intelligence is no longer just a technology story. It has become a contest over political power, economic sovereignty and control of the infrastructure that will underpin the next generation of businesses and public services. Governments that once worried about regulating AI too early are now confronting a different risk: allowing a handful of technology companies to determine how increasingly powerful systems are developed and deployed. Speaking in Madrid on September 21, Spanish Prime Minister Pedro Sánchez argued that the AI industry could not be left to regulate itself. He cited privacy, discrimination, employment, misinformation and the concentration of technological power among the issues requiring government intervention. Spain also outlined a 12-month responsible-AI programme covering cybersecurity, public-interest models, data autonomy and stricter environmental conditions for data infrastructure. The message was larger than Spain. AI is increasingly being treated as strategic infrastructure closer to energy, telecommunications and financial systems than ordinary software. Competing models are taking shape The world is not developing a single system of AI governance. Three broad approaches are emerging. The European Union is building a risk-based regulatory regime through its AI Act. Enforcement powers began applying in August 2026, together with transparency requirements for certain forms of AI-generated content. The European framework places obligations on developers and users according to the potential harm associated with their systems. It addresses documentation, copyright compliance, transparency and the management of risks presented by advanced general-purpose models. Europe’s position is that technological innovation does not remove the state’s responsibility to protect citizens. The United States has taken a more market-led approach. Washington’s policy emphasises innovation, private-sector leadership, national security and competition with China. The US government has also pushed for a uniform national framework rather than a patchwork of different state regulations. That position reflects a wider concern within Washington: strict domestic regulation could weaken American technology companies at a time when AI leadership is becoming a measure of national power. The contradiction is that some of the companies benefiting from this lighter regulatory environment are now calling for government intervention. OpenAI has advocated mandatory safety standards for advanced AI systems, including independent assessments, cybersecurity requirements and incident reporting. It has also called for internationally coordinated technical standards for systems capable of operating with increasing autonomy. The industry still wants room to innovate. But voluntary commitments are becoming harder to defend as the capabilities and economic reach of AI systems expand. America and China are competing and cautiously talking The struggle between the United States and China will shape much of the global AI order. Both countries want control of advanced computing capacity, semiconductor supply chains, intellectual property and international technical standards. Export restrictions and access to sophisticated chips have become instruments of geopolitical strategy. Yet the two powers are also exploring limited cooperation. Recent talks considered an AI incident-notification mechanism and possible protocols covering autonomous systems, malicious non-state actors, cyberattacks and threats to critical infrastructure. This is not evidence of trust. It is recognition that some AI-related failures could damage both countries, regardless of which side leads the technology race. The risk for Africa is that the rules governing advanced AI may eventually be negotiated among powerful states and technology companies, leaving smaller economies to implement standards they had little role in designing. Africa cannot afford to attend these negotiations only after the main decisions have been made. The UN wants a place at the table The United Nations’ Global Dialogue on AI Governance is an attempt to prevent global oversight from becoming an exclusive arrangement among technologically advanced countries. The initiative seeks to bring governments, researchers, civil society and industry into discussions about safety, international cooperation and equitable access to AI. Its legitimacy will depend on whether developing countries receive meaningful influence rather than symbolic representation. African governments will need more than seats at international conferences. They require technically capable delegations that can negotiate questions involving model safety, data governance, intellectual property, computing access and digital trade. Representation without expertise can easily become ceremonial. Africa will also need greater coordination. A fragmented collection of national positions will carry less influence than a common continental approach advanced through the African Union and regional economic communities. Copyright has become an industrial-policy battle The conflict over the material used to train AI models is intensifying. AI developers depend on enormous quantities of text, images, music, audio and video. Creators and publishers increasingly argue that their work should not be used to build commercial systems without consent, attribution or compensation. In Australia, OpenAI and Anthropic have urged the government to reconsider restrictions on using copyrighted local content to train AI models. The companies argue that carefully structured exemptions would encourage investment and technological development. Creative industries fear that such concessions would transfer value from writers, artists, musicians and publishers to foreign technology companies. This debate carries particular importance for Africa. African music, film, photography, journalism, literature, languages and indigenous knowledge are valuable training resources. Without clear rules, they can be absorbed into foreign AI systems without generating corresponding income or ownership for the people who created them. The continent could repeat a familiar economic pattern: exporting raw value and buying back the finished product at a higher price. AI copyright policy is therefore not only a legal matter. It is cultural policy, trade policy and industrial policy. African governments will need positions on licensing, consent, attribution, collective rights management and compensation. Creators should also be represented in the negotiations rather than treated as an afterthought. AI infrastructure comes with physical costs AI may be accessed through a simple application, but the machinery supporting it is resource-intensive. Data centres require electricity, water, land, fibre-optic connections and sophisticated cooling systems. As investment increases, governments are beginning to question who pays for this infrastructure and who receives the economic benefits. Texas recently restricted new data-centre approvals while authorities assessed pressure on the electricity grid and water resources. Regulators were also concerned that speculative connection requests could distort infrastructure planning and pass unnecessary costs to consumers. Africa has an opportunity to attract data-centre and computing investments because of its renewable-energy potential, expanding connectivity and growing digital markets. But governments should avoid agreements that provide cheap power, land, tax incentives and public infrastructure without securing local value. Data centres must not become digital mines, projects that consume African resources while most of the intellectual property and profits are created elsewhere. Agreements should address local employment, skills transfer, electricity obligations, water use, cybersecurity, data jurisdiction and access to computing resources for universities and domestic businesses. The important number should not only be the size of the investment. It should be the amount of productive capacity that remains in the host economy. Jobs will become the most immediate political issue Public anxiety around AI is moving from hypothetical future risks to the immediate effects on employment. AI is already changing customer service, software development, journalism, translation, advertising, design, administration and financial analysis. Entire professions do not have to disappear for the disruption to become significant. Companies may automate routine tasks, reduce entry-level recruitment and expect fewer employees to produce more work. Young people may find that the positions that traditionally provided their first professional experience are precisely those most exposed to automation. African governments cannot respond with general assurances that technology will eventually create new opportunities. They need detailed assessments of which sectors, occupations and communities are most vulnerable. Education and training systems must also move beyond teaching people how to use individual AI applications. Workers need capabilities that remain valuable as the tools change: critical thinking, industry knowledge, communication, data interpretation and the ability to supervise automated systems. An AI strategy without a workforce transition plan is incomplete. Synthetic media is becoming a public-security problem Generative AI has sharply reduced the cost of producing convincing false text, audio, photographs and video. A fabricated recording released shortly before an election could damage a candidate before verification is possible. False statements attributed to a government official could create panic, affect financial markets or worsen diplomatic tensions. Governments will face growing pressure to regulate synthetic political and commercial content. The response must be precise. Poorly drafted misinformation laws can become instruments for suppressing criticism, journalism or satire. Effective regulation should target deceptive conduct and demonstrable harm rather than give public officials broad powers to determine what citizens may say. Electoral commissions, communications regulators, media councils and data-protection authorities will need coordinated systems for authentication, disclosure, evidence preservation and rapid public communication. Citizens must also know when they are communicating with an automated system and when images, voices or videos have been materially generated or manipulated. Africa has a strategy. Implementation is the real test The African Union adopted its Continental Artificial Intelligence Strategy in 2024. It calls for responsible development, stronger research capacity, improved data governance, infrastructure investment and greater African participation in global AI rule-making. The strategy provides a continental foundation. It is not a substitute for national institutions, legislation or budgets. Africa should not copy Europe’s regulatory model wholesale. Nor should it rely entirely on the American belief that markets will resolve most problems. A heavily state-controlled model may also restrict entrepreneurship and public debate. The continent needs a development-focused approach reflecting its economic conditions. Regulation should protect citizens without making compliance impossible for African start-ups. It should encourage investment while demanding local benefits. It should protect personal data while allowing responsible research. It should recognise African languages, cultures and economic priorities rather than treating foreign AI systems as automatically neutral. Governments must pay particular attention when AI is used in policing, recruitment, credit scoring, healthcare, education and public-benefit administration. Decisions in these areas can materially affect a person’s freedom, livelihood or access to essential services. People should be informed when automated systems influence such decisions. They should also have access to a human review and an effective appeal mechanism. Tanzania and East Africa do not need to wait for a crisis East African governments can begin with practical interventions rather than waiting to draft one large AI law. Data-protection authorities can issue guidance on automated processing, model training and biometric information. Communications regulators can establish disclosure expectations for synthetic media and AI-powered customer services. Electoral bodies can introduce standards for artificially generated campaign material. Central banks and financial regulators should examine automated lending, fraud detection, insurance assessment and AI-initiated transactions. Customers need to know when an automated system is making or recommending a financial decision and who carries liability when it causes harm. Public procurement may be one of the strongest immediate policy tools. Suppliers of AI systems to governments should be required to explain what information their systems collect, where the data is stored, how decisions are made and what happens when the technology fails. For Tanzania, the policy conversation should involve the ministry responsible for ICT, the Tanzania Communications Regulatory Authority, the Personal Data Protection Commission, the Bank of Tanzania, electoral institutions, COSTECH and sector regulators. The objective should not be to prevent adoption. It should be to stop public institutions and businesses from acquiring powerful systems without understanding their risks, dependencies and long-term costs. The choice is strategic participation or digital dependency The AI debate is often presented as a choice between rapid innovation and restrictive regulation. Africa’s real choice is between participating strategically and consuming passively. AI could improve agriculture, healthcare, education, financial services, logistics and public administration. It could also expand African creative industries and support the development of services designed around local languages and conditions. None of these outcomes is automatic. Without deliberate policy, AI may deepen dependence on foreign platforms, weaken local creative markets and transfer African data into commercial systems controlled elsewhere. African governments do not need to become hostile to technology. They need to become more sophisticated buyers, regulators, negotiators and investors. The countries that benefit most will not necessarily be those with the strictest laws. They will be those that establish clear national objectives, capable institutions and credible safeguards while creating an environment in which responsible local innovation can grow. Africa should not wait for the United States, China or Europe to finish writing the global AI rulebook. It must enter the negotiations with its interests defined and the institutional capacity to defend them.