Can AI Endorsements Violate Philippine Consumer Law?

Introduction

Artificial intelligence can now generate realistic videos, voices, images, and written testimonials that appear to come from celebrities, influencers, professionals, or ordinary consumers. When a brand uses these materials on social media, websites, livestreams, or online marketplaces, the content may create the impression that a real person personally used, reviewed, or recommended the product.

The legal issue arises when an AI-generated endorsement is presented as genuine without clear disclosure. A fabricated testimonial or celebrity endorsement may constitute a deceptive sales practice or a false, deceptive, or misleading advertisement under Philippine consumer-protection law, particularly when it is intended, or is likely, to induce a purchase.

What Philippine Law Governs AI-Generated Endorsements?

The principal statute is the Consumer Act of the Philippines, or R.A. No. 7394. Its provisions on deceptive sales acts and advertising apply according to the substance and effect of the representation, not merely the technology used to create it.

Article 50 prohibits deceptive acts or practices in connection with a consumer transaction. A representation is deceptive when a producer, manufacturer, supplier, or seller uses concealment, false representation, or fraudulent manipulation to induce a consumer to enter into a transaction. The provision also covers representations that a product has characteristics, benefits, approval, sponsorship, or affiliation that it does not have (R.A. No. 7394; Autozentrum Alabang, Inc. v. Bernardo, et al. (2016)).

Article 110 separately prohibits the dissemination, or causing the dissemination, of a false, deceptive, or misleading advertisement through mail, print, radio, television, outdoor advertising, or any other medium when the material is intended, or likely, to induce the purchase of consumer products or services (R.A. No. 7394).

When Is an AI-Generated Endorsement Deceptive?

An AI-generated endorsement becomes legally problematic when it communicates a material representation that is false or likely to mislead a reasonable consumer. The deception may arise from the words, images, voice, identity, context, or omission of important information.

Examples include the following:

• Creating a video that makes a celebrity appear to recommend a product when the celebrity never made the endorsement;
• Generating a voice clone of a doctor or professional to suggest that the product has been medically tested or approved;
• Publishing fabricated customer reviews that falsely describe personal use or successful results;
• Using an AI-generated influencer who is presented as a real person with genuine experience; and
• Displaying an artificial image or video that implies a sponsorship, affiliation, or approval that does not exist.

Article 50 expressly treats it as deceptive when a seller represents that a product has sponsorship, approval, performance, characteristics, uses, benefits, or affiliation that it does not have. A fake endorsement can fall within these categories if it creates a false impression of approval or personal recommendation.

Why Disclosure May Not Always Cure the Problem

Disclosure is important, but it is not automatically sufficient. A small or unclear disclaimer may not correct a prominent video, headline, voice recording, or testimonial that strongly communicates that a real person endorsed the product.

The assessment considers the advertisement as a whole, including its visual presentation, wording, placement of disclosures, target audience, and the material information omitted. Under Article 110, an advertisement may be misleading not only because of an affirmative false statement but also because it fails to reveal material facts in light of the representations made (R.A. No. 7394).

For example, the statement “AI-generated dramatization” may be insufficient if the advertisement still uses a recognizable celebrity’s face and voice in a manner that suggests actual sponsorship. A clearer disclosure should identify that the person did not make the statement, did not use the product, and did not authorize the content, where those facts are material.

How Courts Assess the Representation

The Consumer Act focuses on whether the representation induces, or is likely to induce, a consumer transaction. Proof that a consumer actually purchased the product is not necessarily required if the advertisement is false, deceptive, or misleading and has the prohibited purpose or tendency.

In Autozentrum Alabang, Inc. v. Bernardo, et al. (2016), the Supreme Court recognized that a deceptive sales act may exist when a seller represents an altered, reconditioned, or second-hand product as brand new. The ruling illustrates that deception may be established by the representation made to the consumer, whether communicated through words or conduct.

The same reasoning is relevant to AI-generated endorsements. A brand cannot avoid responsibility merely because the false statement was produced by software, an advertising agency, a contractor, or an automated content system. The relevant question is whether the business used or caused the use of a misleading representation in connection with a consumer transaction.

Celebrity Images, Voices, and False Affiliation

An unauthorized AI-generated use of a celebrity’s image, name, likeness, or voice may create more than a consumer-protection issue. It may also raise questions involving intellectual property, unfair competition, privacy, contractual rights, and civil liability, depending on the facts.

The Intellectual Property Code prohibits the use of a mark or trade name in commerce when the use is likely to cause confusion, mistake, or deception as to the source, origin, or identity of a business, goods, or services (R.A. No. 8293; Coffee Partners, Inc. v. San Francisco Coffee & Roastery, Inc. (2010)).

Domain names and online commercial identifiers may also create false designation or confusion when they suggest an unauthorized affiliation or sponsorship (Araneta Center Inc. v. Perez (2011); Perez v. Araneta Center, Inc. (2013)). These authorities concern online identifiers and trademarks rather than AI-generated endorsements specifically, but they reinforce the legal importance of preventing false impressions of commercial affiliation.

Liability of Brands and Advertising Participants

Potential responsibility may extend beyond the person who operated the AI tool. Depending on the evidence, it may involve the brand owner, advertiser, advertising agency, production company, influencer intermediary, platform seller, or other participant that created, approved, published, or paid for the misleading material.

Important evidence may include:

• The instructions given to the AI system or production team;
• Approval records, scripts, prompts, and internal communications;
• The identity of the person whose image, voice, or name was used;
• Whether the person authorized the endorsement or actually used the product;
• The placement and wording of disclosures; and
• The relationship between the advertisement and the sale or offer of the product.

A business may not necessarily escape liability by claiming that an employee, agency, or artificial-intelligence tool independently produced the content. The inquiry will ordinarily examine the business’s participation, control, authorization, benefit, and failure to prevent or correct the deceptive communication.

Digital Platforms and Online Enforcement

Online advertisements may be disseminated through social-media accounts, online marketplaces, websites, messaging applications, and livestreaming services. The medium does not remove the material from the Consumer Act. Article 110 expressly covers advertising through “any other medium.”

The Internet Transactions Act of 2023, or R.A. No. 11967, created the E-Commerce Bureau within the Department of Trade and Industry and established an institutional system for monitoring and enforcing compliance in internet transactions. Online businesses should therefore maintain reliable records showing how endorsements were created, approved, disclosed, and removed when challenged.

Where the content involves unauthorized use of a person’s identifiable image, voice, or other personal information, data-protection concerns may also arise. The exact application of the Data Privacy Act will depend on whether personal information is processed, the legal basis for processing, the person’s consent, and the controller or processor’s role.

Possible Regulatory and Civil Consequences

Violations of the Consumer Act may expose the responsible parties to administrative proceedings, corrective measures, penalties, and consumer remedies. In appropriate cases, a consumer may seek rescission and restitution, particularly where the deceptive representation induced the transaction. The Department of Trade and Industry may also address deceptive sales practices under its authority.

In addition, a misleading AI endorsement may support other legal claims, depending on the circumstances:

• Trademark infringement or unfair competition under R.A. No. 8293;
• Violation of contractual or talent-management agreements;
• Claims for damages arising from unauthorized commercial use of a person’s identity;
• Complaints involving misleading online commercial conduct; and
• Privacy or unauthorized-processing claims when personal information is used without a proper legal basis.

These causes of action should not be assumed to apply automatically. The specific identity used, the nature of the advertisement, the authorization given, the product involved, and the evidence of consumer reliance must be examined.

Examples of Higher-Risk Advertising Practices

A beverage company that creates a realistic video of a famous athlete saying that the athlete drinks its product may be engaging in deceptive advertising if the athlete did not record or authorize the statement.

A cosmetics seller that publishes AI-generated “before-and-after” testimonials may create a misleading impression if the depicted customer is fictional, the results are not typical, or the advertisement fails to disclose that the images were artificially generated.

A health-product seller faces greater risk when AI-generated content makes a doctor appear to confirm medical benefits, especially where the product has not been properly evaluated or where the representation may affect a consumer’s health decision.

Recommended Compliance Measures for Brands

Brands using generative AI in advertising should adopt written controls before publication. At a minimum, they should:

• Obtain written authorization before using a person’s name, image, likeness, or voice;
• Verify every factual statement, testimonial, product result, approval, and affiliation;
• Prohibit synthetic endorsements that imply a real person used or recommended a product without proof;
• Use clear and prominent disclosures for fictional, simulated, or AI-generated content;
• Review advertisements for omissions that could materially mislead consumers;
• Keep records of prompts, source materials, approvals, and publication dates; and
• Establish a rapid takedown and correction procedure for disputed or misleading content.

Disclosures should be understandable to the intended audience and placed near the representation they qualify. They should not be hidden in a terms-of-use page, buried among hashtags, or presented in a format that consumers are unlikely to notice.

Practical Legal Test

Before publishing an AI-generated endorsement, a business should ask:

• Would an ordinary consumer believe that a real person made or authorized the endorsement?
• Is the depicted person actually affiliated with, paid by, or authorized by the brand?
• Does the content claim that the person used, tested, or achieved results from the product?
• Are the product’s benefits, quality, approval, or performance accurately represented?
• Is the artificial or simulated nature of the endorsement clearly disclosed?
• Could the content influence a consumer’s purchasing, health, or financial decision?

If the answer to the first question is yes and the underlying endorsement is not genuine, the advertisement presents substantial consumer-protection risk even if the content was created by an automated system or an outside contractor.

Conclusion

Philippine consumer law regulates the message conveyed to consumers, not merely the tool used to produce that message. A fabricated AI testimonial or celebrity endorsement may violate R.A. No. 7394 when it falsely suggests personal experience, sponsorship, approval, affiliation, product performance, or consumer results.

Brands should treat AI-generated endorsements as advertising requiring the same level of substantiation, authorization, review, and disclosure as conventional advertising. The safest approach is to avoid presenting synthetic content as a genuine personal recommendation and to preserve clear records showing that every material claim and identity-based representation is accurate and authorized.

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.