Can AI-Generated Celebrity Ads Violate Philippine Law?
Introduction
Artificial intelligence can now generate realistic videos, voices, images, and statements that appear to come from celebrities or other public figures. When brands use synthetic media in social advertisements without genuine endorsement or clear disclosure, the campaign may create legal exposure even if no real celebrity participated in producing the advertisement.
The principal risks involve deceptive advertising, false claims of sponsorship or affiliation, trademark infringement, unfair competition, and unauthorized processing of personal information. Liability may attach to the brand, advertising agency, platform, or other participants depending on their acts, knowledge, and level of control.
What Makes an AI-Generated Endorsement Legally Risky?
An advertisement becomes legally problematic when it presents a synthetic celebrity image, voice, or statement in a manner that would lead an ordinary consumer to believe that the celebrity actually approved, used, recommended, or was paid to promote the product.
The risk increases when the advertisement uses the celebrity’s name, face, voice, signature mannerisms, prior statements, or recognizable identity without authorization. The same concern applies when artificial intelligence creates a fabricated testimonial about a product’s quality, performance, health benefits, safety, or results.
The fact that the content is computer-generated does not by itself make the advertisement lawful. The legal inquiry generally concerns the message conveyed to consumers, the likelihood of deception, the commercial purpose, and the rights affected.
Consumer Protection Rules on False or Misleading Advertising
R.A. No. 7394, or the Consumer Act of the Philippines, prohibits the dissemination of false, deceptive, or misleading advertisements through Philippine mail or in commerce by print, radio, television, outdoor advertising, or any other medium when the purpose or likely effect is to induce the purchase of consumer products or services.
An advertisement may be misleading not only because of an express false statement, but also because it fails to disclose material facts in light of the representations made. Thus, a social-media advertisement may be deceptive even if every individual word is technically accurate when the overall presentation falsely suggests celebrity approval or product use.
The Consumer Act also identifies deceptive conduct where a seller or supplier falsely represents that a product or service has a sponsorship, approval, performance, characteristic, ingredient, use, or benefit that it does not have. It likewise covers a false representation that the seller or supplier has a sponsorship, approval, or affiliation that it does not possess (R.A. No. 7394).
The Supreme Court has recognized that a seller may commit a deceptive sales act when it represents an altered or second-hand product as brand new, whether the misrepresentation is made through words or conduct (Autozentrum Alabang, Inc. v. Bernardo, G.R. No. 214122, 18 April 2016). The same principle is relevant to synthetic endorsements: the entire presentation, not merely the written caption, may be examined in determining whether consumers were misled.
When Does a Synthetic Endorsement Become Deceptive?
A campaign is particularly vulnerable when it communicates any of the following false impressions:
- Actual endorsement: the celebrity personally recommends the product;
- Product use: the celebrity has used the product and achieved the advertised result;
- Commercial relationship: the celebrity is an ambassador, sponsor, investor, or business partner;
- Professional approval: the celebrity or an apparent expert has evaluated and approved the product; or
- Authentic testimonial: the words, voice, or video are genuine statements by the person portrayed.
A disclaimer may reduce confusion in some circumstances, but a small, obscure, or difficult-to-read disclaimer may not cure a dominant and realistic representation that the celebrity genuinely endorsed the product. Disclosure should be prominent, intelligible, and placed where consumers will see it before relying on the advertisement.
Trademark Infringement and Unauthorized Commercial Use
Artificial intelligence-generated advertising may also infringe trademark rights when it uses a registered mark, trade name, or a confusingly similar sign in commerce without authorization. Section 155 of the Intellectual Property Code, or R.A. No. 8293, covers unauthorized use of a reproduction, counterfeit, copy, or colorable imitation of a registered mark in connection with the sale, offering for sale, distribution, or advertising of goods or services when the use is likely to cause confusion, mistake, or deception.
Infringement may occur upon the prohibited commercial use even without proof of an actual completed sale. The statutory rule recognizes liability when the infringing material is used in connection with advertising or other commercial activity and is likely to cause confusion (Ginebra San Miguel, Inc. v. Tanduay Distillers, Inc., G.R. Nos. 196372, 210224, 216104, and 219632, 14 November 2022).
The relevant question is not limited to whether the AI-generated material copies a registered logo. A campaign may create liability when it uses a mark, business name, or distinctive branding in a manner that suggests affiliation, sponsorship, or origin that does not exist.
Trade names also receive protection even before or without registration against unlawful acts, including subsequent use by another party that is likely to mislead the public (R.A. No. 8293, Section 165.2; Coffee Partners, Inc. v. San Francisco Coffee & Roastery, Inc., G.R. No. 169504, 24 March 2010).
Unfair Competition and Passing Off
Even when a brand does not copy a registered trademark, the use of a celebrity’s identity or a competitor’s presentation may support an unfair competition claim if it causes consumers to believe that the advertised goods come from, or are connected with, another person or business.
Section 168.3 of R.A. No. 8293 treats as unfair competition conduct that gives goods the general appearance of another manufacturer’s or dealer’s goods, or otherwise clothes them with an appearance likely to deceive the public and defraud another of legitimate trade.
The Supreme Court has explained that the assessment is based on the general impression made on the ordinary purchaser, who may act with carelessness and haste, rather than on an artificial side-by-side comparison of every feature (Del Monte Corporation v. Court of Appeals, G.R. No. 78325, 25 January 1990). A highly realistic deepfake may therefore be assessed by its overall commercial impression.
Use of a Celebrity’s Image, Voice, and Personal Information
An AI-generated endorsement may involve the collection, reproduction, alteration, or dissemination of information relating to an identifiable individual. Depending on the circumstances, the person’s photograph, voice, name, contact details, account information, or other identifying material may constitute personal information subject to the Data Privacy Act.
The National Privacy Commission has treated the continued online posting of a former agent’s photograph and personal information in advertisements after withdrawal of consent as unauthorized processing under the Data Privacy Act. The agency also emphasized that financial involvement or organizational association, without substantial evidence of direct participation, does not automatically establish liability (NPC 20-026, JBA v. FNT and NNT, 2022).
This does not mean that every AI-generated advertisement automatically violates the Data Privacy Act. The analysis depends on the information used, the lawful basis for processing, the person’s consent or other applicable justification, the notice given, the purpose of processing, and the parties that controlled or participated in the publication.
Separate civil claims may also arise from unauthorized commercial exploitation of a person’s identity, particularly where the content causes reputational, financial, or professional harm. The available authorities do not establish a single general statutory rule that automatically resolves every unauthorized AI likeness claim; the facts and the legal theory pleaded remain important.
Liability of Brands, Agencies, and Platforms
The brand is ordinarily the party with the greatest exposure when it commissioned, approved, paid for, or benefited from the advertisement. It may be difficult for a brand to avoid responsibility merely by asserting that an outside agency or software provider generated the content.
An advertising agency may face separate exposure when it created, edited, approved, distributed, or knowingly maintained the misleading material. Contractual allocation of responsibility may support indemnification between the parties, but it does not necessarily eliminate liability to consumers, rights holders, or regulators.
Platform liability depends on the platform’s role, knowledge, participation, and the specific law or regulatory order invoked. The Internet Transactions Act of 2023, R.A. No. 11967, created the E-Commerce Bureau under the Department of Trade and Industry and provides a regulatory structure for monitoring and enforcing compliance in internet transactions. The Act should be reviewed together with its implementing rules and any applicable administrative orders before relying on a particular enforcement mechanism.
Possible Legal Consequences
Depending on the facts, affected parties may pursue civil, administrative, or criminal remedies. Potential consequences include removal or correction of advertisements, administrative penalties, damages, injunctions, trademark remedies, unfair competition actions, and consumer redress.
Consumer-protection proceedings may focus on whether the advertisement was false, deceptive, or materially misleading. Intellectual-property proceedings may examine unauthorized use and likelihood of confusion. Privacy proceedings may focus on the processing and dissemination of personal information without a lawful basis or beyond the scope of consent.
These causes of action may overlap, but they do not have identical elements. A brand should therefore avoid assuming that defeating a trademark claim necessarily defeats a consumer-protection or privacy complaint.
Common Examples
Fabricated product testimonial. A company creates a video showing a celebrity apparently saying that a supplement cured a medical condition. This may raise consumer-protection concerns because it falsely represents the product’s benefits and the celebrity’s approval.
Voice-cloned endorsement. A brand uses an AI-generated voice that closely resembles a celebrity’s voice and places it in a paid social-media advertisement. Liability may arise if the presentation suggests an actual endorsement or uses personal attributes without authorization.
Fake investment promotion. An advertisement depicts a well-known public figure recommending an investment platform. The campaign may mislead consumers about sponsorship, affiliation, or approval even if the celebrity’s name is not formally registered as a trademark.
Clearly labeled synthetic advertisement. A company uses an artificial character and clearly states that the character is fictional and computer-generated. The risk of deception may be lower, although the campaign must still avoid false product claims, unauthorized use of protected marks, and unlawful processing of personal information.
Compliance Measures Before Publishing an AI Advertisement
- Obtain written authorization. The agreement should expressly cover synthetic images, voice cloning, digital replicas, editing, paid advertising, social-media publication, geographic scope, duration, and post-termination use.
- Verify every factual claim. Substantiate statements concerning product performance, health benefits, quality, safety, price, availability, and consumer results.
- Disclose synthetic content. If artificial media is used, provide a prominent disclosure that is understandable to the target audience and visible before the consumer relies on the message.
- Review the entire advertisement. Examine the visual material, voice, caption, hashtags, landing page, comments, and call-to-action for an overall misleading impression.
- Document approvals. Preserve the source files, prompts, licenses, contracts, substantiation records, disclosure versions, approval history, and takedown procedures.
- Control continued publication. Remove or update advertisements when consent expires, the endorsement relationship ends, the product changes, or a claim is found to be inaccurate.
Conclusion
AI-generated endorsements are not unlawful merely because artificial intelligence was used. The legal risk arises when synthetic media falsely suggests that a celebrity endorsed, used, approved, or is affiliated with a product, or when the campaign uses protected identity, branding, or personal information without proper authority.
Brands should treat an AI-generated endorsement as a regulated commercial communication, not merely as creative content. Before publication, they should secure express rights, verify all claims, disclose synthetic elements, assess trademark and privacy implications, and maintain records showing responsible review and control.
About Nicolas and De Vega Law Offices
Nicolas and de Vega Law Offices is a full-service law firm in the Philippines. You may visit us at the 16th Flr., Suite 1607 AIC Burgundy Empire Tower, ADB Ave., Ortigas Center, 1605 Pasig City, Metro Manila, Philippines. You may also call us at +632 84706126, +632 84706130, +632 84016392 or e-mail us at [email protected]. Visit our website https://ndvlaw.com.
