- Introduction
The emergence of deepfakes has introduced a new challenge for Nigerian law as artificial intelligence (AI) now has the ability to misrepresent a person's identity through fabricated digital content that looks genuine. A notable example occurred on 27 May 2026, when an audio recording attributed to President Bola Ahmed Tinubu circulated online and was subsequently reported to have been AI-generated. The recording allegedly captured the President making inflammatory remarks about insecurity, foreign loans, and the 2027 general election. Although the recording was later identified as AI-generated, it was convincing enough to generate rapid reactions, including a threat of prosecution against a commentator who had circulated it, hours before the clip's origin was even verified.2 This problem is not limited to politics. In June 2025, the Advertising Regulatory Council of Nigeria (ARCON) investigated advertisements that used the image and voice of Pastor Enoch Adeboye and Channels Television journalist, Seun Okinbaloye, to promote medical products. ARCON's preliminary findings indicated that the material was AI-generated and that neither individual had consented to it.3 Similarly, in February 2026, Ngozi Okonjo-Iweala warned Nigerians about an AI-manipulated video that falsely used her image to promote an investment scheme promising unrealistic returns.4 Each of these incidents shows the same pattern. A person's identity can be separated from their actual conduct and used to manufacture credibility for a product, message or transaction.
At the centre of these developments is therefore a question of digital identity. When technology can reproduce a person's face, voice or conduct convincingly, what legal protections exist against the creation and use of a digital representation that the individual did not authorise? Nigeria currently has no law written specifically for deepfakes. Existing laws can nevertheless address many of the harms deepfakes may cause. The difficulty is that these laws were largely designed around the underlying harm rather than the manipulation of digital identity itself. This article examines how the existing Nigerian legal framework applies to deepfakes, where that framework becomes uncertain, and whether Nigeria's emerging AI governance structure is capable of addressing those gaps.
2. The Existing Legal Framework
There is currently no legislation that specifically addresses deepfakes in Nigeria. In the absence of a dedicated statutory framework, a regulatory framework could be pieced together from existing legal provisions that were developed before the widespread use of generative artificial intelligence (AI).
2.1 Cybercrimes (Prohibition, Prevention, etc.) Act 2015 (as amended)
The Cybercrimes Act 5 represents one of the principal statutes through which certain forms of deepfake-related misconduct may be addressed. These include identity theft and impersonation,6 unauthorised access to computer systems, offensive or obscene communications, and cyberstalking.7 It also covers messages sent with intent to cause annoyance, injury, or needless anxiety,8 unauthorised interception of communications>,9 and imposes data-retention obligations on service providers>.10 Of these provisions, section 22 is the most relevant since it addresses identity theft and impersonation. The section originally criminalised identity theft and impersonation committed by employees using their specialised knowledge within financial institutions. However, with the 2024 amendment, the provision now applies to all public and private organisations. While the amendment was not drafted with deepfakes in mind, it may nevertheless apply where AI-generated content is used to impersonate a person in connection with the services of a public or private organisation, provided the other elements of the offence are established.
These sections can, in theory, apply to the incidents above. In practice, their application still depends on proving that a specific individual behind a synthetic voice or face intended to defraud a specific victim. This is difficult to establish. By the time the fraud is discovered, the trail is already several steps removed from any real person a court can question.
2.2 Nigeria Data Protection Act 2023
The Nigeria Data Protection Act (NDPA) does not expressly regulate deepfakes, but its provisions on personal data may extend to synthetic representations of identifiable persons. A person's face and voice are personal data, and where they reveal biometric characteristics, they may qualify as sensitive personal data requiring a higher standard of protection and consent.11 Processing of such data must comply with principles including lawfulness, fairness, purpose limitation and accuracy.12 The creation of a deepfake that misappropriates a person's likeness without consent or a lawful basis may therefore amount to a data protection concern in its own right. This means that liability may arise from the unauthorised processing of the individual’s data or personal attributes even where the deepfake has not yet been used to defraud, defame or otherwise harm the person.
2.3 Constitution of the Federal Republic of Nigeria 1999 (as amended)
The Constitution protects rights that may be directly implicated by deepfake technology. These include the right to the dignity of the human person,13 the right to private and family life,14 and the right to freedom of expression and information.15 A victim of a defamatory or degrading deepfake therefore has a legitimate claim to dignity and privacy, while the creator or distributor may claim free expression. This creates a constitutional tension that Nigerian case law has not yet addressed specifically in relation to synthetic media.
2.4 Criminal Code,16 Penal Code,17 and sector regulators
The legal consequences of a deepfake may also be addressed under general criminal and civil laws, as well as sector-specific regulation, depending on the nature of its use. A false or degrading representation of an individual may engage criminal provisions relating to defamation or obscenity under the Criminal Code (in the southern states)18 and the Penal Code (in the north),<19 in addition to the civil tort of defamation. A deepfake that falsely portrays a public figure as endorsing a product, service, or investment may also fall within the jurisdiction of the Advertising Regulatory Council of Nigeria (ARCON) in relation to misleading advertising. The growing use of AI-generated impersonation in fraudulent schemes further demonstrates the risk associated with this technology. In February 2026, an individual was arraigned over an alleged ₦230 million fraud involving an AI-cloned image of Peter Obi,20 while BUA Group subsequently disowned an AI-generated video falsely portraying its Executive Chairman as endorsing an investment scheme<.21
3. Where the Existing Framework Falls Short
The existing legal framework reveals three significant gaps. First, none of these statutes distinguishes AI-generated content from other forms of manipulated media. As a result, liability generally depends on proving subjective intent to defraud, defame, or harass, rather than the creation or dissemination of synthetic content itself. Second, the existing framework provides limited guidance on questions of authenticity and proof when synthetic media is presented as evidence. The Tinubu audio incident exposed this uncertainty when the Presidency called for prosecution hours before the clip's origin was verified.22 Third, the investigation and adjudication of deepfake-related cases may require technical expertise that existing enforcement institutions and courts have limited experience in applying.
The financial sector shows how this gap is currently being addressed in a piecemeal manner. The Central Bank of Nigeria's March 2026 Anti-Money Laundering (AML) directive requires banks, fintechs and mobile money operators to strengthen their fraud-detection systems using measures such as behavioural analysis and machine-learning tools.23 This may improve the ability of regulated institutions to detect synthetic-identity fraud, but it remains a sector-specific compliance response. It does not establish a general rule on the admissibility or reliability of AI-generated evidence, it does not determine the remedies available to victims or provide a comprehensive framework for allocating responsibility when fraudsters defeat an institution's verification controls. These issues should also be distinguished from the existing rules on electronic evidence. Section 84 of the Evidence Act 2011 governs the admissibility of electronic and computer-generated evidence in judicial proceedings,24 and in Kubor v Dickson25 the Supreme Court held that a party relying on such evidence must satisfy the conditions in section 84(2), whether by certificate or by oral evidence. Those conditions are directed at the proper functioning of the computer and the regular supply of information to it, not at whether a recording shows what it claims to show. The Evidence (Amendment) Act 2023 subsequently introduced sections 84A to 84D, addressing related matters including information in electronic form, computer-generated records and authentication of electronic records.26 However, these provisions do not expressly contemplate the particular authentication and evidential challenges associated with AI-generated content and deepfakes.
The same fragmentation can be observed in the political context. Commentary on the 2027 election cycle has raised questions about the absence of clear rules separating the creation of manipulated political material from its subsequent dissemination.27 Taken together, these gaps show that Nigeria's response is developing incrementally across different offences, sectors and institutions. What is missing is a framework that treats synthetic media as a distinct legal and evidential problem rather than addressing its consequences one at a time.
4. Where Regulation is Headed
Nigeria's AI governance framework remains a work in progress. The National Artificial Intelligence Strategy, published in September 2025 under the Federal Ministry of Communications, Innovation and Digital Economy with NITDA as a key implementing body, sets out a five-year vision for AI and identifies governance as one of its core pillars.28 The institutional structure required to give effect to this vision, however, remain under discussion. Policy analysts have advocated for an AI Governance Regulatory Body comprising representatives from academia, civil society, and industry to provide oversight of NITDA's proposed Code of Practice.29 As at the date of this article, the proposed body has not yet been constituted and the Code of Practice remains unfinalised. Organisations and individuals should therefore expect further developments before the regulatory position becomes fully settled.
At the same time, Nigerian regulators have demonstrated a willingness to exercise their statutory powers against technology companies without waiting for legislation directed specifically at the technology concerned. The Federal Competition and Consumer Protection Commission's $220 million fine against Meta and WhatsApp was upheld on appeal in April 2025.30 However, the Nigeria Data Protection Commission's (NDPC) separate fine against Meta arising from the same broader investigation was eventually resolved by consent judgement without payment.31 The contrasting outcomes show that regulatory power on paper does not guarantee a uniform enforcement outcome in practice. For deepfake cases, where questions of authenticity, attribution and intent are already harder to prove, this inconsistency is likely to matter even more. The effectiveness of enforcement may depend not only on the existence of regulatory powers but also on the clarity and coordination of the institutions exercising them.
5. What this Means in Practice
The absence of a specific deepfake law does not leave victims without legal protection. Depending on how the content was created or used, existing laws on impersonation, data protection, defamation, fraud and misleading advertising may apply. For organisations, this means internal verification processes matter more than ever. An AI-generated voice or video of an executive can be used to instruct staff to release payments or share confidential information, and the instructions can sound or look completely genuine. Organisations should therefore make verification routine. No payment or disclosure should follow from a call, video or message alone. Unusual or high value transfers, such as those to a new payee or involving changed bank details, should require a second approver. Urgent instructions should also be confirmed through a separate channel that the requester did not choose, for example the executive’s office line from the company directory, the executive’s assistant, or in person. Pressure to act quickly or to keep the request secret should itself be treated as a warning sign.
Proving authenticity is also becoming harder. The BUA Group case demonstrates this well. Before the company could deal with the reputational damage, it first had to publicly confirm that the video was fake. Nigerian law already has rules on electronic evidence, but sophisticated synthetic media raises new questions about how authenticity should be proven in court. Media, advertising, and political organisations should build reliable content verification practices now, ahead of the 2027 election, rather than wait until misinformation has already caused harm.
6. Comparative Regulatory Approaches
Other jurisdictions have already begun addressing different parts of the problem, and their approaches provide useful starting points for Nigeria.
The European Union (EU) has approached the issue through transparency. Article 50(4) of the AI Act, which has applied since 2 August 2026, requires deployers of AI systems that generate or manipulate images, audio or video content constituting a deepfake to disclose that the content has been artificially generated or manipulated.32 A separate transition period, 2 December 2026, applies to the machine-readable marking duty on providers of generative AI systems already on the market. This rule does not determine liability, but it creates a transparency requirement that can later assist a court or regulator. Nigeria currently has no equivalent rule. In a dispute, therefore, there may be no comparable disclosure trail for a court or regulator to consider.
South Korea provides a more targeted response to the electoral risks of deepfakes. Article 82-8 of its Public Official Election Act prohibits the production, editing or distribution of deepfake campaign content during the 90 days preceding an election. Outside that period, AI-generated campaign content must be clearly labelled, and violations may attract up to seven years' imprisonment or a fine.33 More importantly, South Korea has also invested in the ability to enforce these rules. Its National Forensic Service, working with the Ministry of the Interior and Safety, developed an AI deepfake analysis model, which the Ministry says contributed to the detection and deletion of more than 10,000 pieces of illegal deepfakes election material during the 2025 presidential election period.34 This is relevant for Nigeria ahead of the 2027 elections. A rule prohibiting political deepfakes would have limited value if the institutions responsible for enforcing it lack the technical capacity to identify manipulated content.
Even where a party claims that a video, image or audio recording is a deepfake, how does the court determine whether that claim is correct? The United States has been considering this problem at the federal level. Its Advisory Committee on Evidence Rules, meeting in Washington on 7 May 2026, declined to advance a proposed Rule 707 on machine-generated evidence and declined to publish a draft Rule 901(c) addressing deepfakes, with some members questioning whether a special rule is needed at all.35 Louisiana took a different route. Act 250, which came into force on 1 August 2025, requires lawyers in civil proceedings to exercise reasonable diligence to verify the authenticity of evidence before offering it, treats the offering of evidence that the lawyer knew or should have known was false or artificially manipulated, without disclosing that fact, as a violation, and provides a pretrial process for raising authenticity objections.36
The United Kingdom (UK) provides another useful comparison because it draws a line between sharing deepfake content and creating it. Section 188 of the Online Safety Act 2023 criminalised sharing, or threatening to share, intimate images without consent, including images that appear to show a person, which brings deepfakes within scope.37 The Data (Use and Access) Act 2025 subsequently criminalised creating, or requesting the creation of, a purported intimate image of an adult without consent or reasonable belief in consent from 6 February 2026, even where the content is never shared.38
Denmark presents a fifth and more unconventional approach by treating personal likeness as something capable of attracting property-like protection.39 Its proposed amendment to the Copyright Act would extend copyright-style protection to a person's face, body and voice, requiring consent before a realistic digital replica can be made available to the public.40 The proposal also contemplates remedies similar to those available for copyright infringement and protection extending for fifty years after death. The first version of the bill lapsed when Denmark held a general election.41 The bill has since been reintroduced in the Danish Parliament with a proposed start date of 1 January 2027, and as at the date of this article it has not been enacted.42 If enacted, its significance would extend beyond deepfakes themselves. It would allow the harm to be framed not only as misuse of personal data or fraud, but as the unauthorised use of an aspect of a person's identity. It should be noted, however, that although the bill sits within the Copyright Act, it is not intended to create copyright in the strict sense.<43
These examples show that Nigeria's response does not have to begin and end with a new deepfake offence. The EU shows how disclosure can create transparency. South Korea demonstrates the importance of combining legislation with detection capability. Louisiana addresses the evidentiary problem through professional diligence. The UK demonstrates why the creation of harmful deepfakes may need to be treated separately from their distribution. Denmark offers a possible intellectual property route for protecting personal likeness. Together, these approaches provide a menu of reforms that could address the different gaps currently left open by Nigeria's legal framework. None of them, however, has fully solved the problem of proving authenticity or reliably detecting synthetic content and on the evidential question the United States has so far chosen to wait at the federal level. They should therefore be treated as options to adapt, not finished answers.
Adapting these options to Nigeria raises three practical questions. The first is constitutional. A rule modelled on South Korea's election period ban would restrict political expression, and section 39 of the Constitution protects freedom of expression subject to section 45(1), under which a restriction must be contained in a law that is reasonably justifiable in a democratic society for a recognised purpose, such as public order or the protection of the rights of others.44 A Nigerian provision would therefore need to be narrowly drawn, limited to a defined pre-election period and to content realistic enough to deceive voters, and supported by a labelling duty outside that period, as the Korean law itself provides. The second question is institutional. A Danish-style likeness right depends on civil remedies and court orders, including orders for the removal of unlawful content from the platform where it appears,45 but a victim of a fast-moving deepfake needs removal before litigation concludes. That points to a regulator with takedown powers, and NITDA and the NDPC are the most likely candidates, although their coordination in this area is untested. The third question is proof. The EU disclosure duty and the Korean detection model both assume technical capacity that Nigerian courts and enforcement agencies are still developing, so a reform that creates new duties without investing in forensic capability would be difficult to enforce.
7. Conclusion
Is Nigerian law ready for deepfakes? Not entirely. Liability exists, but it is scattered across laws that were never written for this purpose and is dependent on proving intent and origin, both of which synthetic media makes hard to establish. Closing this gap requires Nigeria’s AI governance structure, which is still not fully in place, to treat synthetic media as its own legal issue rather than a version of harms the law already knows. The comparison above suggests an order of priority. The first step is proof, an amendment to the Evidence Act dealing expressly with the authentication of recordings alleged to be AI-generated, together with investment in forensic capacity, because every other remedy depends on being able to show what is real. The second is a narrowly drawn, time limited rule on deceptive synthetic political content, paired with labelling duties, ideally in time for the 2027 elections. Disclosure duties for AI providers and a statutory right in personal likeness are worth considering only once the institutional framework is settled. Until then, victims, organisations and regulators are left relying on a patchwork of older laws. The question is no longer whether deepfakes are a legal problem in Nigeria. It is how quickly the law can catch up with them.
Footnotes
1 Mary Ojo, Trainee Associate, Intellectual Property & Technology Department, S.P.A. Ajibade & Co., Lagos Office, Nigeria.
2 Nanji Nandang, 'Presidency threatens legal action against VDM over alleged Tinubu audio' (27 May 2026) available at (https://www.tv360nigeria.com/presidency-threatens-legal-action-against-vdm-over-alleged-ai-generated-tinubu-audio/) ICIR (27 May 2026) accessed 22nd August 2026.
3 Jeremiah Agada, ‘ARCON Raises Alarm Over Fake Ads Featuring Pastor Adeboye, Seun Okinbaloye, Other, Threatens Legal Action Against Meta, Advertisers’ Brandcom available at( https://brandcom.ng/2025/06/13/arcon-raises-alarm-over-fake-ads-featuring-pastor-adeboye- seun-okinbaloye-other-threatens-legal-action-against-meta-advertisers/) accessed 21st September 2026.
4 Christabel Ndoeche, ‘Okonjo-Iweala warns Nigerians against fake AI investment video’ The Punch (11 February 2026) available at (https://punchng.com/okonjo-iweala-warns-nigerians-against-fake-ai-investment-video/) accessed 22nd September 2026.
5 See, the Cybercrimes (Prohibition, Prevention, etc) Act 2015, as amended by the Cybercrimes (Prohibition, Prevention, etc) (Amendment) Act 2024 (Cybercrimes Act).
6 The Cybercrimes Act, s22.
7 The Cybercrimes Act, s24.
8 The Cybercrimes Act s24(1)(b).
9 The Cybercrimes Act s12.
10 The Cybercrimes Act ss37-38.
11 The Nigeria Data Protection Act 2023 (NDPA 2023) s30.
12 NDPA 2023 s24.
13 The Constitution of the Federal Republic of Nigeria 1999 (CFRN 1999) s34.
14 CFRN 1999 s37.
15 CFRN 1999 s39.
16 The Criminal Code Act 2004.
17 The Penal Code Act 2004.
18 The Criminal Code Act, ss373-375 (defamation) and ss233B-233D (obscene publications).
19 The Penal Code, ss391-395 (defamation) and s200 (obscene or indecent acts).
20 Wale Igbintade, ‘Court Remands Businessman Accused of Using AI to Impersonate Peter Obi in N230m Forex Fraud’ ThisDay available at (https://www.thisdaylive.com/2026/02/19/court-remands-businessman-accused-of-using-ai-to-impersonate-peter-obi-in-n230m-forex-fraud-2/) accessed 22nd September 2026.
21 Christabel Ndoeche, ‘BUA disowns AI investment video, warns public’ The Punch (8 June 2026) available at (https://punchng.com/bua-disowns-ai-investment-video-warns-public/) accessed 22nd September 2026.
22 Oludare Ogunlana, ‘The 2027 Nigerian Election and the AI Swarm Threat’ OSRS (27 May 2026) available at (https://www.ogunsecurity.com/post/the-2027-nigerian-election-and-the-ai-swarm-threat) accessed 22nd September 2026.
23 10 March 2026 CBN Circular, Issuance of Baseline Standards for Automated Anti-Moneylaundering (AML) Solution for Financial Institutions in Nigeria available at
(https://www.cbn.gov.ng/Out/2026/CCD/CBN%20issues%20Baseline%20Standards%20for%20A utomated%20Anti-Money%20Laundering%20Solution.pdf) accessed 22nd September 2026.
24 The Evidence Act 2011, s84.
25 Kubor & Anor v. Dickson & Ors (2013) 4 NWLR (Pt. 1345) 534.
26 The Evidence (Amendment) Act 2023, ss84A-84D.
27 Femi Aderibigbe, ‘Nigeria Wants to Fight AI Deepfakes Before 2027. The VDM Controversy Shows the Rules Are Still Missing’ Development Diarie (30 May 2026) available at
(https://developmentdiaries.com/nigeria-wants-to-fight-ai-deepfakes-before-2027-the-vdm-controversy-shows-the-rules-are-still-missing/) accessed 22nd August 2026.
28 Federal Ministry of Communications, Innovation and Digital Economy, summarised in OECD.AI, ‘National Artificial Intelligence Strategy (NAIS)’ available at (https://oecd.ai/en/dashboards/policy-initiatives/national-artificial-intelligence-strategy-nais) accessed 22nd August 2026.
29 Tunani, Policy Brief: Nigeria’s National AI Strategy available at (https://tunanihq.org/assets/publication/Policy-Brief-NNAIS-Strategy-final.pdf) accessed 22nd August 2026.
30 Federal Competition and Consumer Protection Commission, ‘Violations: Tribunal Upholds FCCPC’s $220 Million Fine Against Meta/WhatsApp’ (25 April 2025) available at
(https://fccpc.gov.ng/violations-tribunal-upholds-fccpcs-220-million-fine-against-meta-whatsapp/) accessed 22nd August 2026.
31 Ndidiamaka Ede, ‘$32.8 Million Fine Dispute: Meta Settles with Nigerian Govt Out of Court’
Premium Times (3 November, 2025) available at (https://www.premiumtimesng.com/news/headlines/832580-32-8-million-fine-dispute-meta-settles-with-nigerian-govt-out-of-court.html) accessed 22nd September 2026.
32 Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act) [2024] OJ L, 2024/1689, 12.7.2024, as amended, arts 50(2), 50(4) and 113 (consolidated version of 27 July 2026) available at(https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX%3A02024R1689-20260727) accessed 10th October 2026.
33 Public Official Election Act (Republic of Korea), arts 82-8 (unofficial English text) available at (https://koreanics.com/laws/public-official-election-act/article-82-8/) accessed 11th October 2026.
34 Kim Tae-kyung 'Successful analysis of 15 cases in the first half of the year, support for deletion of over 10,000 illegal election materials through cooperation with the National Election Commission’ Financial News (30 July 2025) available at (https://en.fnnews.com/news/202507301001446959) accessed 10th October 2026.
35 Nete Raymond 'US Judicial panel delays action on AI-generated evidence, deep fakes' Reuters (7 May 2026) available at (https://www.reuters.com/legal/government/us-judicial-panel-delays-action-on-ai-generated-evidence-deep-fakes-2026-05-07/) accessed 10th October 2026; Advisory Committee on Evidence Rules, Report to the Committee on Rules of Practice and Procedure (17 May 2026), in Committee on Rules of Practice and Procedure, Agenda Book (3–4 June 2026)
(https://www.uscourts.gov/sites/default/files/document/advisory_committee_on_evidence_rules_ may_2026.pdf) accessed 15th October 2026.
36 Louisiana Act No 250 of 2025 (HB 178), amending Louisiana Code of Civil Procedure arts 371 and 1551 (in force 1 August 2025).
37 Sexual Offences Act 2003, s66B, inserted by Online Safety Act 2023, s188 available at (https://www.legislation.gov.uk/ukpga/2023/50/section/188) accessed 12th October 2026.
38 Data (Use and Access) Act 2025, s138 (inserting ss66E–66H into the Sexual Offences Act 2003) available at (https://www.legislation.gov.uk/ukpga/2025/18/section/138) accessed 12th October 2026.
39 Sofia Karttunen, ‘The Danish approach to copyright and deepfakes: A model for the EU?’ European Parliamentary Research Service, p2 available at (https://www.europarl.europa.eu/RegData/etudes/ATAG/2026/782611/EPRS_ATA(2026)782611_EN.pdf) accessed 12th October 2026.
40 Karttunen (n39) p1; Luis Rijo, ‘Denmark’s deepfake bill makes publishers prove consent from January 2027’ PPC Land (9 October 2026) available at (https://ppc.land/denmarks-deepfake-bill-makes-publishers-prove-consent-from-january-2027/) accessed 12th October 2026.
41 Olivier Lombardo, ‘A new sense of self: Denmark’s copyright amendment against deepfakes’ Dennemeyer Blog (16 February 2026) available at (https://www.dennemeyer.com/blog/posts/a-new-sense-of-self-denmark-s-copyright-amendment-against-deepfakes) accessed 12th October 2026.
42 Rijo (n40); Roselyne Min, ‘Denmark introduces bill to protect people’s faces and voices from deepfakes’Euronews (8 October 2026) available at (https://www.euronews.com/2026/10/08/denmark-introduces-a-bill-protecting-peoples-faces-and-voices-from-deepfakes) accessed 12th October 2026.
43 Karttunen (n39).
44 CFRN 1999 ss39 and 45(1).
45 Rijo (n40).
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.






