Nigeria, South Africa and Kenya are introducing measures to tackle AI-generated political misinformation ahead of upcoming elections, while governments across the continent develop sector-specific rules to address the risks associated with artificial intelligence.
African countries are stepping up efforts to regulate artificial intelligence and protect their democratic institutions against the growing threat of AI-generated misinformation, deepfakes and synthetic media.
Recent developments across Nigeria, South Africa, Kenya and Ghana indicate that governments are increasingly relying on existing regulatory institutions, electoral commissions and sector-specific policies to address the risks associated with the rapidly expanding use of artificial intelligence.
The developments come as AI-generated images, videos, cloned voices and automated content become increasingly accessible, raising concerns about their potential misuse in political campaigns, electoral processes, financial services and public administration.
According to an October 2026 review published by Africa’s AI Almanac on November 8, African AI governance is increasingly taking shape through existing institutions responsible for competition, consumer protection, financial regulation, data privacy and elections.
The review highlights a growing shift from broad discussions about responsible AI towards practical regulatory measures designed to address specific risks in different sectors.
Among the countries making notable progress are Nigeria, South Africa, Kenya and Ghana, although their approaches and levels of implementation differ.
Nigeria establishes AI unit to tackle electoral misinformation ahead of 2027 elections
Nigeria is among the countries taking direct steps to address the potential misuse of artificial intelligence during elections.
The Independent National Electoral Commission (INEC) has established a dedicated Artificial Intelligence unit within its Information and Communication Technology Directorate to monitor digital threats, detect deepfakes and counter attempts to manipulate electoral information ahead of the country’s 2027 general elections.
INEC Chairman, Professor Joash Amupitan, announced the development on October 14, 2026, while delivering a keynote address at the Artificial Intelligence and Democracy Forum in Abuja.
The forum, organised in collaboration with Yiaga Africa, focused on safeguarding Nigeria’s 2027 elections in the age of artificial intelligence.
Amupitan warned that AI-generated videos, cloned audio recordings and other forms of synthetic media could be used to impersonate political candidates and electoral officials, circulate fabricated election guidelines, announce unauthorised results and discourage citizens from participating in elections.
To address these threats, the commission plans to deploy digital monitoring tools, deepfake detection software and data analytics to identify suspicious content and attempts to manipulate electoral information.
INEC is also developing a digital strategy to monitor online information and working towards collaboration with technology platforms to facilitate the identification and labelling of AI-generated electoral content.
The commission has further indicated that artificial intelligence could support election administration through logistics planning, voter education, security monitoring and the identification of inconsistencies in uploaded election results.
However, Amupitan stressed that important electoral decisions would remain under human control.
“Algorithms will not, and must never, decide the 2027 vote,” he said.
The chairman maintained that human accountability, statutory oversight and judicial review would remain central to the electoral process.
The development is particularly significant for Nigeria, where social media platforms have become major channels for political communication, campaign activities and the distribution of election-related information.
As the 2027 elections approach, the ability of electoral authorities to identify manipulated content and respond quickly to misinformation could become an important part of protecting public confidence in the democratic process.
South Africa tests disinformation rules amid AI-generated political content
South Africa has also strengthened its approach to election-related misinformation through the introduction of an election Disinformation Code.
The Electoral Commission of South Africa implemented the code on September 4, 2026, establishing rules intended to address the spread of misleading information during election campaigns.
The framework places responsibilities on political parties and candidates regarding election-related communication and seeks to address the circulation of deceptive content that could undermine the integrity of the electoral process.
The country’s November 4 local government elections provided an early test of the measures.
During the final week of campaigning, the Electoral Commission flagged an AI-generated video featuring former South African president Nelson Mandela being used for political messaging.
The incident highlighted the growing challenge of synthetic media, particularly when fabricated content uses the likeness or voice of respected historical figures to influence public opinion.
By that stage, the commission had received more than 700 complaints relating to election misinformation, with most of the reported cases linked to content circulating on X.
The development demonstrates how AI-generated political content can complicate efforts by electoral authorities to distinguish authentic information from manipulated material.
South Africa’s approach is notable because it places specific responsibilities on political parties and candidates rather than relying exclusively on electoral authorities to identify and remove misleading content.
Kenya prepares deepfake detection measures ahead of 2027 general elections
Kenya is also preparing to strengthen its electoral safeguards against AI-generated misinformation ahead of its 2027 general elections.
The Kenyan government plans to introduce artificial intelligence-powered deepfake detection tools six months before the election, following growing concerns about the potential misuse of synthetic audio, manipulated videos and automated political content.
The National Development Implementation Committee has identified AI-generated deepfakes and synthetic audio as major threats to the country’s upcoming electoral process.
Francis Odhiambo, a commissioner at the Independent Electoral and Boundaries Commission (IEBC), has also ranked the misuse of artificial intelligence among the significant threats facing the 2027 elections.
Concerns about AI-generated political content in Kenya have increased following findings in a September 2026 threat report by AI company Anthropic.
The report documented the activities of an actor who used Anthropic’s Claude AI model to produce political content at scale targeting Kenyan audiences.
The incident illustrates how generative AI tools can be used to produce and distribute politically themed material rapidly, potentially making it more difficult for citizens to distinguish genuine political communication from coordinated manipulation.
Kenya’s planned deployment of detection tools is therefore intended to strengthen the country’s ability to identify synthetic media before it spreads widely during the election period.
However, the effectiveness of such technology will depend on the availability of reliable verification systems, trained personnel and cooperation between electoral authorities, social media companies and independent fact-checkers.
Ghana introduces additional safeguards for AI adoption
Beyond election security, Ghana is strengthening its approach to artificial intelligence regulation, particularly in financial services and public institutions.
The Bank of Ghana is developing a directive to guide the use of AI in the financial sector, with particular attention to data quality, model risk, cybersecurity and consumer protection.
The proposed framework reflects concerns about the increasing use of automated systems in financial services, including credit assessment, fraud detection, customer management and other banking operations.
Ghana is also introducing additional oversight for AI systems acquired by public institutions.
Communications Minister Samuel Nartey George announced a directive requiring AI systems purchased by government institutions to undergo centralised technical reviews.
The assessments are expected to examine interoperability, data protection, cybersecurity and local skills transfer.
The initiative is intended to ensure that public institutions adopt AI systems that meet appropriate technical and security requirements while supporting the development of local technological capacity.
These measures form part of Ghana’s wider efforts to translate its AI ambitions into practical governance arrangements.
Nigeria moves towards sector-specific AI regulation
Nigeria’s regulatory activities are also extending beyond election security.
The Federal Competition and Consumer Protection Commission (FCCPC) has released draft Sales Promotion Regulations addressing the use of AI and automated tools in promotions targeting Nigerian consumers.
Under the proposed rules, businesses using AI in promotional activities would be required to disclose such use, clearly label AI-generated content, avoid manipulative practices and provide consumers with an option to opt out of automated marketing.
The proposed regulations also seek to prevent businesses from avoiding responsibility for harmful outcomes simply because an automated system was involved.
Meanwhile, the Nigeria Data Protection Commission has called for privacy-by-design principles to be incorporated into the country’s approach to regulating advanced AI technologies.
The commission has recommended integrating ISO/IEC 42001:2023, an international standard for AI management systems, into Nigeria’s proposed AI Regulatory Sandboxing Platform.
In the healthcare sector, the government is also working towards sector-specific data regulations addressing the use of clinical information, anonymisation and patient consent.
These developments suggest that Nigeria’s AI governance framework is gradually expanding through existing regulatory institutions while broader AI governance arrangements continue to take shape.
Kenya and South Africa reconsider broader AI regulatory frameworks
While several African countries are introducing sector-specific safeguards, efforts to establish comprehensive national AI legislation are progressing at different speeds.
In Kenya, the Senate ICT Committee has recommended that the Senate should not proceed with the proposed Artificial Intelligence Bill, 2026, in its current form.
The committee argued that the proposed legislation was too broad and potentially too restrictive for the country’s current stage of AI development.
It recommended that the government first complete its national AI policy before proceeding with comprehensive legislation.
South Africa is also revising its proposed national AI policy following problems identified in an earlier draft, including references to sources that could not be verified.
An independent panel chaired by Wits University researcher Benjamin Rosman is advising on the revision.
The revised policy is expected to reach Cabinet in March 2027 before proceeding towards public consultation. A final approval date has not yet been established.
These developments demonstrate that African governments are still working through how to balance innovation, public protection and the institutional capacity required to enforce AI regulations.
African Union calls for stronger continental AI governance
At the continental level, the African Union’s Continental Artificial Intelligence Strategy, adopted by the AU Executive Council in July 2024, provides a common framework for developing responsible AI across member states.
The strategy promotes a people-centred and development-oriented approach, with emphasis on building African AI capabilities, strengthening cooperation, encouraging investment and addressing potential risks.
It also recognises challenges associated with algorithmic bias, discrimination, data protection, disinformation, surveillance and the impact of AI on vulnerable communities.
More recently, participants at e-Gov’A 2026 adopted the Déclaration de Yaoundé on October 17, encouraging the responsible use of AI in public services.
The declaration calls for public-sector AI systems to be transparent, secure, fair and accountable, with human oversight, auditing mechanisms and opportunities for individuals to appeal decisions that significantly affect their rights.
It also provides for the establishment of an Observatory of e-Government and AI in Africa to monitor progress.
In another development, the African Union has called for technology-facilitated gender-based violence and AI-related threats to be incorporated into Women, Peace and Security National Action Plans and early-warning systems.
The call reflects growing concerns about how digital technologies and AI-generated content can be weaponised to harass women, discourage their participation in public affairs and undermine peacebuilding processes.
The challenge of implementing AI regulations across Africa
Despite the progress being made, significant challenges remain in translating AI policies and regulatory proposals into effective enforcement.
A major concern is the difference in institutional capacity across African countries.
Many governments are developing AI strategies and introducing new rules without having sufficient technical expertise, funding and infrastructure to monitor increasingly complex AI systems.
An October 2026 analysis by the Institute for Security Studies (ISS) highlighted this implementation gap, noting that 44 African countries have adopted data protection legislation, while only 38 have operational data protection authorities.
This demonstrates that passing legislation does not automatically guarantee effective enforcement.
Regulators also need specialists who understand how AI systems operate, how algorithms make decisions and how to assess potential risks involving discrimination, privacy, security and transparency.
Another challenge is the cross-border nature of digital misinformation.
AI-generated content can be created in one country and distributed across several others within minutes through social media platforms and messaging applications.
This makes cooperation between governments, electoral commissions, technology companies, journalists and fact-checking organisations increasingly important.
For elections in particular, detection technology alone may not be sufficient. Electoral institutions also need transparent communication systems, rapid fact-checking mechanisms, public awareness campaigns and clear procedures for addressing manipulated content without restricting legitimate political expression.
Conclusion: Africa must balance AI innovation with public protection
The latest developments indicate that Africa’s AI governance landscape is moving towards more practical and sector-specific regulation.
Nigeria is preparing electoral safeguards ahead of its 2027 general elections, South Africa has tested election disinformation rules, Kenya is planning to deploy deepfake detection tools, and Ghana is strengthening oversight of AI adoption in financial services and public institutions.
However, the effectiveness of these initiatives will depend on whether governments can move beyond policy announcements to establish institutions capable of implementing and enforcing the rules.
As artificial intelligence becomes increasingly integrated into elections, public services, financial systems and everyday communication, African countries face the challenge of protecting citizens without unnecessarily restricting innovation.
Ultimately, the goal should not simply be to regulate artificial intelligence, but to build an environment where the technology can support economic development, strengthen democratic institutions and serve the public interest while remaining accountable to the people it affects.
