Authored By: Treasure Walter
Babcock University
INTRODUCTION
With the rapid advancement and growing adoption of Artificial Intelligence (AI), the world is undergoing a significant transformation in the way we live, work, communicate, and conduct business. Take for example, scrolling through TikTok or Instagram and coming across an AI-generated virtual influencer advertising a skincare brand with over one million followers. She promotes the product, offers skincare advice, and even claims that she personally uses and recommends the brand. Now, imagine that something goes wrong. A report emerges that about 30 people who used the product experienced serious adverse reactions. At this point, an important question arises, who is to be sued, the AI system itself, the company that built it, or the brand that deployed it to sell the product
To understand this thesis, there must be comprehension of the existing legal framework. This paper argues that Nigerian and the international law currently lack a clear answer, because both framework were built to regulate real people, not artificial ones.[1]
CONCEPTUAL FRAMEWORK
First of all, there is a difference between a virtual influencer and a deep fake. A virtual influencer is a computer-generated digital character used on social media to post content, interact with followers, and promote brands just like a real person. Virtual influencers can be divided into human influencers, CGI/fictional influencers, AI-generated influencers and fully autonomous persons.[2] Section 2 of the TAKE IT DOWN ACT[3] defines deep fake also known as digital forgery, as an artificial intelligence (AI) generated or manipulated image, video, or audio file that appears completely real and indistinguishable from an authentic recording to a reasonable observer.[4]
Brands use virtual influencers to promote their business because it’s faster and saves the energy of going around the world to do it manually. They use AI-generated influencers because they are both time-efficient and cost-effective. Engaging human influencers often involves significant expenses, which can be challenging for emerging brands, whereas AI-generated influencers provide a more affordable alternative while still achieving marketing objectives.[5] A virtual influencer can have an enormous economic presence without having a legal personality.[6]
NIGERIA’S LEGAL FRAMEWORK: ARCON AND FCCPA
In Nigeria’s response to the growing influence of digital advertising, came in the form of the Advertising Regulatory Council of Nigeria Act 2022 was introduced, which establishes ARCON as the apex responsible body for regulating advertising content across every platform, including social media. This was to ensure transparent digital marketing endorsement. What does ARCON do? They review and approve advertisements before they are published, broadcast, or run online via its Advertising Standards Panel to ensure they are ethical, legal, and decent. If you run an ad without ARCON approval or certificate, there will be penalties attached. The Lagos Division of the Federal High Court, in the case of Digi Bay Limited (Trading under the name and style of Betway Nigeria) & 2 others v Attorney General of the Federation & Anor[7], in the presence of Honourable Justice Akintayo Aluko, allows ARCON to regulate every form of advertisement in Nigeria.[8] The ARCON Act 2022 included digital marketing as part of the regulations. From 2025, ARCON mandated that all digital adverts must be approved before exposure; no advertisement should be made without the approved signature of the Director General. Now the important issue is whether the ARCON Act was enacted with only humans in mind or whether there is a clause that allows for the use of artificial intelligence. ARCON replaced APCON to oversee traditional human and corporate endorsements. Still, it now focuses on digital marketing, such as AI-generated marketing, to reduce the rate of people investing in Ponzi schemes or falling prey to fraudulent advertising.[9]
Beyond ARCON, the Federal Competition and Consumer Protection Act 2018 offers a second, sharper tool: section 126, which prohibits false, misleading, or deceptive endorsement, including undisclosed sponsorships and unauthorized testimonials. Section 126 of the Federal Competition and Consumer Protection Act 2018 states “An undertaking shall not, for the purpose of promoting, directly or indirectly, the supply or use of any goods or services or any business interest, make a representation to the public that a test as to the performance, efficacy or length of life of the goods or services has been made by any person or publish a testimonial with respect to the products, unless it can establish the matters specified in subsection (2)…”[10] the issue then becomes, can a virtual influencer, incapable of using or experiencing any product, ever give a testimonial within the meaning the drafters intended?
Therefore, a close reading of both instrument reveals a striking silence, neither the ARCON act nor the FCCPA defines or even mention, an artificial or non-human endorsement. Nigerian law regulates who paid for the ad and what claims were made, but never asks whether the person making the claim exist at all. This set out the gap my paper is built to expose.
INTERNATIONAL/ COMPARATIVE FRAMEWORK
Where Nigeria stays silent, other countries have started speaking. Under the FTC’s revised Endorsement Guides(2023) of the United States, “endorser” is expressly redefined as the “party whose opinions, beliefs, findings or experience the message appears to reflect, which could appear to be an individual, group of people, or institution”. It includes writers of fake reviews and non-existent entities that give endorsements.[11] This brings virtual influencers in scope; failure to disclose AI/synthetic origin is treated as deceptive under the FTC Act s.5. If a brand is to advertise using an AI-generated influencer, they are to state that it is AI-generated that way to prevent confusion.
The European system is fully against deepfakes or digital forgery in order to reduce deception and manipulation. The system introduced the AI Act, Article 50, which is to take effect from 2 August 2026; the Act was provided to detect and label deepfake or manipulated AI videos. While the Code of Practice is voluntary, adherence to AI Act Article 50 is compulsory. Scholars argue that virtual influencers within this scope lack legal personhood and therefore cannot be traders under the UCPD (Unfair Commercial Practices Directive), thereby pushing liability to the deployer brand instead. This is a persuasive analysis, as AI-generated influencers lack physical bodies to test a skin product and therefore cannot have a reaction to a skin product, leaving the liability to the brand behind the AI-generated image. This is genuinely useful comparative contrast to Nigeria’s advertisement-focused model.
The UK does not have a single, standalone law governing deepfakes or AI-generated influencers but instead regulates it through customer protection like the Advertising Standards Authority (ASA) and the sister of the law that accompanies it, Committee of Advertising Practice (CAP)[12]. These laws are independent regulators that regulates advertising online, they provide mechanism in which people can advertise product but under safety care. There is also the financial conduct authority which ensures that marketing communication should state the nature of the contract offered if not it will attract penalties and this self-regulatory framework helps in promoting that all marketing structures are well conducted.
CROSS-CUTTING LEGAL ISSUES
Scholars argue against AI having legal personalities; that AI is not real and therefore they can’t be sued or have personal liability; instead, those liabilities fall on the company that operates that AI (developer, deployer, brand).
In the Intellectual Property sector, which comprises copyright, trademark and patent has their own laws in Nigeria that guide personal ownership and have now evolved the use of AI-generated content. IP raises a parallel ownership problem which is, who owns the virtual influencer? For example, a company creates Lil Miquela; who owns Lil Miquela? A writer who edits and redefines AI-assisted text retains authorship because human creative input shaped the final work. Section 28(1) of Nigeria’s Copyright Act 2022,[13] states “except as otherwise provided in an agreement, copyright conferred by this Act shall initially vest in the author”. And even when work is done collectively, the work is still owned by the author, as seen in section 29(a)(b) of Nigeria’s Copyright Act 2022[14]. Copyright covers matters that deal with books, poems, songs, etc. Since the law states that who owns the work is the author or the one who created it, if a poet who loves to make poems uses ChatGPT in composing his poem, who owns the work? Is it ChatGPT that gave the idea or the poet that made the poem? According to section 28 of the copyright act, the poet is the owner of the work because he composed it but the twist is that if the means by which the poem was composed was done directly by the AI, no editing or addition of point of any sort, then the work cannot have copyright. The UK and US copyright regimes similarly require human authorship and creative input for a work to qualify for protection.
Trademark laws raise a different question from copyright: not who created the persona but who owns the brand asset and who is exposed when the asset has been misused. Trademark protects your brand name, logo, and proprietary services from imitation and online infringement. Whether the virtual influencer’s name/face can be trademarked as a brand asset, and the infringement risk if her design too closely resembles a real person’s likeness, trademark protects AI from using a person’s logo, brand, etc. Nigerian trademark law protects distinctive marks used in trade regardless of whether the mark represents a real or fictional mascot. Nigerian law has no dedicated right of publicity statute and would likely rely on passing off to protect a real individual.
RECOMMENDATION
ARCON should amend the Nigeria Code of advertising to something clear and more pronounced, conspicuous label should be more transparent wherever an endorser is AI-generated or synthetic so as to provide clear standards. The EU AI Act should be reviewed and use to examine digital marketing in different states so as to enlighten ignorant people on these mechanism. The Federal Competition and Consumer Protection Act 2018 should be amended to redefine ‘endorsement’ and ‘testimonial’ under section 126[15] so as to examine and capture representations made by non-human personas. At present, section 126 defines ‘testimonials’ as “a person or any person who has tested a product”. That statement should be more open so as to adopt new means of testing product, the law cannot fight what does not exist so if the statement accommodates entities with no physical appearances, then the law can fight it when it breaches any legal regulation. The United States offers a reasonable workable template: The Federal Trade Commission’s Revised Guide[16] which defines an endorser as the party whose opinions, beliefs, findings or experience the message appears to reflect, which could appear to be an individual, group of people, or institution. The definition drafted broadly enough to capture endorsers that do not exist in person, therefore adapting this language to section 126[17] of the FCCPA 2018 will resolve the most difficult question of AI legal personhood.
Secondly, Nigerian law should codify the developer, deployer and the brand liability directly, rather than leaving it to be inferred. The prevailing scholarly position is AI system should not be granted legal personality since they cannot bear right and obligation because there are fictional unlike a juristic or a natural person. Since an AI system cannot bear such rights and obligation, the liability should be attached to the human or the operating system behind them that is the developer, the deployer or even the brand that the AI system is advertising for.¹ The European system houses the strongest backing law in relation to AI system. The European Union’s Artificial Intelligence Act reflects this logically, distinguishing between the providers who build the AI systems and the deployer who put the system to use, placing a distinct obligation on each.² [18]Nigeria does not have any equivalent statutory allocating liability under the FCCPA that currently attaches to any undertaking or person acting on its behalf,³ which is broad enough to reach a brand but does not clearly resolve disputes between a company, brand, its marketing agency, and the AI developer where several parties are involved. The Nigerian system dedicates deployer-liability provision, which names the deploying brand as the primary bearer of the liability for representations made through an AI-generated persona. This will would remove this ambiguity entirely.
CONCLUSION
In conclusion, EU article 50 is a binding legislation with specific disclosure duty which does not rely on interpretation but target the problem directly making it the strongest law on AI system. The US FTC’s redefinition of an ‘endorser’ is more precise and directly closes the gap that Nigeria has but the endorsement guides are not standalone statute but rather there are enforcement guidelines which makes them strong but not as strong as the EU and Nigeria ARCON and the FCCPA are statutorily strong for advertising generally but only on specific issues and not accommodating the full problem. The UK regulatory law is the weakest on paper but pragmatic. It is self-regulatory, penalties are not fully statutory but reputational. There are all sufficient laws but have different hierarchy of strength and effectiveness.
Bibliography
Primary sources
Advertising Regulatory Council of Nigeria (ARCON) Act 2022.
Federal Competition and Consumer Protection Act 2018, s 126.
Nigerian Code of Advertising Practice, Sales Promotion and Other Rights/Restrictions on Practice 2021 (ARCON).
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, art 50.
Digi Bay Ltd & 2 Others v Attorney General of the Federation & Anor (Unreported, Suit No FHC/L/CS/1262/2024, Federal High Court, 29 April 2025).
Secondary sources
Abdu-Salaam Abbas & Co, ‘The Powers of the Advertising Regulatory Council of Nigeria to Regulate Social Media Posts: A Review of the Federal High Court’s Decision in Digi Bay Limited & 2 Others v Attorney General of the Federation & Anor’ (Abdu-Salaam Abbas & Co) https://abdu-salaamabbasandco.com/the-powers-of-the-advertising-regulatory-council-of-nigeria-to-regulate-social-media-posts-a-review-of-the-federal-high-courts-decision-in-digi-bay-limited-2-others-v-attorney-general-of/ accessed 3 September 2026.
Advertising Standards Authority, ASA System Response to CMA Consultation on Unfair Commercial Practices Guidance (22 January 2025) https://assets.publishing.service.gov.uk/media/67ee8f439eae202448299c95/Advertising_Standards_Authority.pdf accessed 3 September 2026.
European Commission, ‘Code of Practice on Transparency of AI-generated Content’ (Shaping Europe’s Digital Future, last updated 31 July 2026) https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content accessed 3 September 2026.
Nwafor Chinecherem, ‘What Are the Benefits of Having an ARCON Certificate? The Uses’ (SoniBaze Digital, 18 June 2026) https://sonibaze.ng/benefits-of-having-arcon-certificate-nigeria/ accessed 2 September 2026.
[1] Kiruthiga P, ‘The Legal Implications of Virtual Influencers and AI Personas on Social Media Platform: A Comparative Analysis of Global Jurisdictions’ (2025) 7(1) Indian Journal of Law and Legal Research
[2] Mauro Conti, Jenil Gathani and Pier Paolo Tricomi, ‘Virtual Influencers in Online Social Media’ (2022) 60(8) IEEE Communications Magazine 86.
[3] TAKE IT DOWN Act, s 2
[4] TAKE IT DOWN Act, Pub L No 119-12, 139 Stat 58 (2025), s 2, inserting Communications Act of 1934, s 223(h)(1)(B) (47 USC § 223).
[5] Frédéric Bossard and Florence Ciais-Bossard, ‘Who Will Trust Virtual Influencers?’ (SKEMA Knowledge, 26 May 2025) https://knowledge.skema.edu/virtual-influencers-trust-ethics-marketing-brands/ accessed 1 September 2026.
[6] ‘The Perils and Promises of Virtual Influencers: Where EU Consumer Law Falls Short’ (Oxford Law Blogs, 15 November 2024) https://blogs.law.ox.ac.uk/oblb/blog-post/2024/11/perils-and-promises-virtual-influencers-where-eu-consumer-law-falls-short accessed 1 September 2026.
[7] Digi Bay Ltd & 2 Others v Attorney General of the Federation & Anor (Unreported, Suit No FHC/L/CS/1262/2024, Federal High Court, 29 April 2025).
[8] Nwafor Chinecherem, ‘What Are the Benefits of Having an ARCON Certificate? The Uses’ (SoniBaze Digital, 18 June 2026) https://sonibaze.ng/benefits-of-having-arcon-certificate-nigeria/ accessed 2 September 2026.
[9] ‘Stop using Tinubu’s image for AI-generated adverts, ARCON warns’ The Guardian Nigeria accessed 2 September 2026.
[10] ‘Section 126 Federal Competition and Consumer Protection Act 2018’ LawGlobal Hub accessed 2 September 2026.
[11] Douglas Thompson, ‘Truth or Consequences: FTC Revised Endorsement Guides & Class Action Risk’ accessed 2 September 2026.
[12] Advertising Standards Authority, ASA System Response to CMA Consultation on Unfair Commercial Practices Guidance (22 January 2025) https://assets.publishing.service.gov.uk/media/67ee8f439eae202448299c95/Advertising_Standards_Authority.pdf accessed 3 September 2026.
[13] Copyright Act 2022, s 28(1).
[14] Copyright Act 2022, s 29(a)-(b).
[15] Federal Competition And Consumer Protection Act 2018, S 126.
[16] Federal Trade Commission, Guides Concerning the Use of Endorsements and Testimonials in Advertising, 16 CFR part 255 (revised 2023).
[17] ibid
[18] Artificial Intelligence Art (EU) 2024/1689, arts 3(3)-(4) (definitions of ‘provider’ and ‘deployer’)
