Compliance
On 21 September 2026 the Ad Standards Council’s circular on artificial intelligence in advertising took effect. It does not ban anything, and it does not make an AI-made ad suspect. What it does is add a disclosure and a named responsibility to work that many Philippine brands are already running. If your approval chain does not have a line for either, this is the week to add one.
What the circular actually says
ASC Circular 2026-014, Guidelines on the Use of Artificial Intelligence in Advertising, was published on 18 September and took effect on 21 September 2026. Three things in it change how an ad gets approved.
→ It applies to everyone in the chain. The Circular covers “advertisers, creative agencies, production companies, media agencies, digital platforms, and all other persons or entities involved in the creation, production, publication, or dissemination of advertising materials that utilize AI, whether in whole or in part”. Not just the advertiser. Not just the studio.
→ AI is not the problem; silence is. “The use of AI, by itself, shall not render an advertisement misleading, deceptive, or otherwise noncompliant.” But it “shall not exempt an advertiser, advertising agency, or any other responsible party from compliance” with everything that already applied.
→ Disclosure is now required. Advertising “created, generated, or materially enhanced” through AI, “whether in whole or in part”, must carry “a clear and reasonably conspicuous disclosure informing the audience that AI was used”.
What the disclosure has to look like
The Circular names acceptable wording and leaves room for equivalents: “AI-generated,” “AI-assisted,” “Created with AI,” “Contains AI-generated content,” “or words of similar import that clearly communicate the use of AI to the audience”.
On placement it is equally plain. The disclosure must be “reasonably noticeable, legible, and appropriate to the format of the advertisement whether in font size or as seen against a background”. A four-point line over a moving background is not a disclosure. Neither is one that only appears in the caption when the asset gets reposted without it.
Decide the wording once, per campaign
Pick the phrase, the size and the placement at the brief stage, and put it in the deliverables list alongside the aspect ratios. Retrofitting a legible disclosure onto forty finished cut-downs is a re-export of every one of them.
The line that decides who is accountable
Section V is the part worth reading to the client. Advertisers “shall exercise appropriate human oversight and professional judgment in the use of AI”, and they “remain fully responsible for all claims, representations, depictions, endorsements, and other content appearing in advertisements, regardless of whether such content was created by humans, AI, or a combination of both”.
Then the sentence that ends the argument: “Reliance on AI shall not excuse or mitigate any violation of applicable laws, rules, regulations, or advertising standards.”
In an approval chain that means the sign-off cannot be delegated to the tool, and “the model produced it” is not a defence anyone can raise later.
Four more things the Circular points at
It also lists circumstances where other law already demands disclosure, consent or authorisation, and says those continue to apply.
| What it names | What it means on a production |
| Minors | Rules on “the participation, representation, or portrayal of minor-aged talents in advertising” apply to generated children as well as real ones |
| Identifiable people | Content that depicts or is “likely to be mistaken for an identifiable real person, including public figures or celebrities” can engage publicity, personality and privacy rights, and may need that person’s consent |
| Intellectual property | Copyright, trademarks, patents and industrial designs apply to what a model generates. Who owns the output is a separate question — see who owns AI-generated content in the Philippines |
| Buildings and protected property | Rights, licences and permissions still govern “the use or depiction of buildings, structures, artworks, or other protected properties” |
The platforms have their own rules, and they are not the same
Clearing the ASC does not clear YouTube. Google requires creators “to disclose when they use AI to meaningfully alter or generate photorealistic content”, and lists three triggers: content that “makes a real person appear to say or do something they didn’t do”, “alters footage of a real event or place”, or “generates a realistic scene that didn’t actually occur”. Selecting that setting labels the video for viewers.
Meta works the other way round. For its ads products it says that where its generative tools make “a significant edit to the image or video, we will apply an AI label either behind the three-dot menu, or next to the Sponsored label”, and that where the result includes “an AI-generated photorealistic human, the label will appear next to the Sponsored label”. One platform asks you to declare; the other decides and labels for you.
So a single spot can need an on-screen disclosure for the ASC and a platform-level flag on upload. They are two different jobs, done by two different people, and neither one covers the other.
Where this goes in the approval chain
→ At brief. Is AI being used, and in which parts? Record it. The answer decides everything downstream.
→ At concept. Any identifiable person, any minor, any building, any borrowed style. Flag each one for a rights check before it is generated, not after it is loved.
→ At first cut. Disclosure wording, size and placement approved as a deliverable, in every ratio.
→ At delivery. A named person signs that the claims are accurate and substantiated. One person, by name, not “the team”.
→ At upload. Platform-level AI flags set per channel, by whoever publishes.
Keep the production record
The Circular expects human oversight and professional judgment. The evidence of both is the brief, the direction given, the versions rejected and the checks run. Build that record as the work is produced; it is also what settles a complaint quickly.
General information, not legal advice. Read the Circular in full and have your own counsel advise on your campaigns.
If a complaint lands
The ASC “may require the advertiser, advertising agency, or other responsible party to address or resolve the matter with the complainant”, and nothing in the Circular stops it “from issuing a Cease and Desist Order (CDO) or taking such other action as may be warranted under its rules”. The cost of a missing disclosure is not a fine so much as a stopped campaign.
Where PUNX fits
PUNX produces AI-generated brand content from Makati — 50-plus people, founded 2023, for Procter & Gamble, San Miguel, Petron, GCash and Sun Life. Concept and direction are set by people, generation runs on an AI pipeline, and the disclosure decision is part of the deliverables list rather than a question asked at delivery. Brand-safe AI production covers the provenance side.
FAQ
Do Philippine ads have to disclose the use of AI?
Yes. ASC Circular 2026-014, effective 21 September 2026, requires advertising created, generated or materially enhanced through AI, whether in whole or in part, to carry a clear and reasonably conspicuous disclosure telling the audience that AI was used.
What wording counts as an AI disclosure?
The Circular names “AI-generated,” “AI-assisted,” “Created with AI,” “Contains AI-generated content,” or words of similar import that clearly communicate the use of AI. It must be reasonably noticeable, legible and appropriate to the format, in font size and against its background.
Who is responsible if an AI-made ad breaches the rules?
The advertiser and the other responsible parties. The Circular says they remain fully responsible for all claims, representations, depictions and endorsements regardless of whether the content was made by humans, AI or both, and that reliance on AI does not excuse or mitigate a violation.
Is an ASC disclosure enough for YouTube and social?
No. Platforms run their own rules. YouTube requires creators to disclose when AI meaningfully alters or generates photorealistic content, including a real person appearing to say something they did not, altered footage of a real event, or a realistic scene that never occurred. That is a separate flag set at upload.
Related
AI content production — the service, from brief to delivery
Brand-safe AI production — disclosure, consent and the record behind them
Who owns AI-generated content — what the IP Code actually says
Campaign in production right now?
Send us the cut and where it runs. We will tell you what has to be disclosed, where it goes, and who signs it.
Start a project ↗