# How to label AI-generated ads on Meta > When Meta adds an AI info label, when you must disclose AI in ads, and what ASCI's new labelling guidelines ask of Indian advertisers. A worked example. - **Author**: Marxx - **Published**: 2026-10-31 - **Category**: AI - **URL**: https://marxx.ai/library-guides/how-to-label-ai-generated-ads-on-meta --- On Meta, an "AI info" label can be added to your ad by Meta, and you must disclose AI-made media yourself in social issue, election and political ads. In India, ASCI's new guidelines also ask you to add your own label when AI-made content could change a buyer's decision. All facts below were checked on 2026-10-20. This is not legal advice. ## What Meta does Meta says ad images created or significantly edited with its generative AI creative tools may show "AI info" in the **About this ad** screen in the three-dot menu. If the image has an AI-generated photorealistic human, the label sits next to the Sponsored label. Minor edits such as resizing or colour correction get no label ([Meta Help Centre](https://www.meta.com/en-gb/help/artificial-intelligence/355108217670024/), [Meta newsroom](https://about.fb.com/news/2025/02/gen-ai-transparency-metas-ads-products/)). From 1 June 2026 Meta also uses automated detection to find ad media made or edited with third-party AI tools. No advertiser action is needed for that ([Meta Transparency Center](https://transparency.meta.com/policies/ad-standards/SIEP-advertising/SIEP/)). ## What you must disclose to Meta If your ad is about social issues, elections or politics, Meta requires you to disclose a photorealistic image, video or realistic audio made or edited with third-party generative AI when it shows: - a real person saying or doing something they did not, - a realistic person who does not exist, or an event that did not happen, or - altered footage of a real event. If you do not disclose, Meta says it will reject the ad, and repeated failure may bring penalties. For ordinary commercial ads, the Meta pages we checked ask for no disclosure from you. ## What ASCI asks in India ASCI released its guidelines on synthetically generated content on 29 September 2026. They take effect three months after publication ([ASCI press release](https://www.ascionline.in/wp-content/uploads/2026/09/Press-Release_Guidelines-for-Responsible-Labelling-of-Synthetically-Generated-Content-in-Advertising.pdf), [guidelines](https://www.ascionline.in/wp-content/uploads/2026/09/Guidelines-for-Responsible-Labelling-of-Synthetically-Generated-Content-in-Advertising.pdf)). | Case | ASCI rule | |---|---| | Fake endorsements or testimonials, exaggerated results, non-existent places shown as real, unauthorised copyrighted work, deepfakes, a person's likeness without consent | Prohibited, even with a label | | AI influencers or ambassadors, a real person's face or voice copied for personalised messages, even with consent, fabricated events or settings, demos of products that do not exist yet, AI sound effects that sell a product's core feature | Label required | | Colour correction, lighting, noise removal, decorative backgrounds, obvious fantasy effects, ad copy, subtitles and translations | No label needed | Labels can read "Audio/Video created using AI" or "Audio/Video enhanced using AI". Platform labels are allowed. The rules apply when the content could mislead a buyer. India's IT Rules were also amended on 10 February 2026. The government says platforms must ensure clear labels and traceable metadata for permitted AI content ([PIB, 6 August 2026](https://www.pib.gov.in/PressReleasePage.aspx?PRID=2295500®=1&lang=1)). That duty sits with platforms. ## Steps 1. List every ad that uses AI-made images, video, voice or avatars. 2. Ask which row of the table each one falls in. 3. Delete anything in the prohibited row. 4. Add a short on-screen label to ads in the label row, such as "Created using AI". 5. For political ads, use Meta's disclosure. 6. Keep a record of the tool and the date. ## A worked example Say you make three ads for a INR 899 serum. - **Ad 1:** a real founder, with AI-made Hindi subtitles. No label needed, since subtitles and translations are exempt. - **Ad 2:** an AI presenter called "Riya" who explains how to use the serum. Label it, because ASCI lists synthetic influencers. - **Ad 3:** an AI-made before and after of a "customer". Do not run it. A fabricated testimonial is prohibited even with a label. Read [how to avoid AI-looking ads](/guides/how-to-avoid-ai-looking-ads), [what AI UGC is](/glossary/ai-ugc) and [AI UGC vs human UGC](/compare/ai-ugc-vs-human-ugc). For the tools Meta launched, see [Meta's Advertising Week 2026 AI tools](/news/meta-advertising-week-2026-ai-ad-tools). ## FAQ ### Do I have to label every AI ad on Meta? Not by ticking a box on ordinary ads, as far as Meta's pages show. Meta may add its own label. Political ads need your disclosure. ### Does an AI label make a misleading ad safe? No. ASCI says some uses are prohibited regardless of any label. ### Does ASCI apply to ads made before the date it takes effect? ASCI says the guidelines take effect three months after publication. Plan to label new work from then, and check older ads you keep running. --- - [More AI articles](https://marxx.ai/library-guides/category/ai) - [All articles](https://marxx.ai/library-guides)