AI Content Policy
Last updated
In short: we make video using AI, and we are open about it. Indian rules now require some AI advertising to be labelled, and ban certain uses outright. This page explains which is which, what we label, and what we will not make. We supply the disclosure wording with every delivery.
We make video ads using AI. We think that is a good thing, because it makes production faster and far cheaper. But it comes with rules, and those rules changed in 2026. This page explains how we work within them, so you know what you are getting and what you need to do at your end.
The rules that apply
Two things changed for AI-generated advertising in India:
- The Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026, in force from 20 February 2026, require synthetically generated content to be labelled and to carry provenance information where technically possible.
- The Advertising Standards Council of India (ASCI) issued its final Guidelines for Responsible Labelling of Synthetically Generated Content in Advertising, dated 17 September 2026. They take effect three months after publication, in December 2026, and set out when an ad that uses AI must be labelled and what is not allowed at all.
We already work to these rules, ahead of the date they take effect. The ASCI Code that sits behind them applies to every ad today, and building compliance in costs far less than fixing it after a complaint.
The three categories
ASCI sorts AI use in advertising by whether it could mislead a consumer. The category decides whether a label is needed.
| Category | What it covers | What happens |
|---|---|---|
| Prohibited | Fake endorsements or testimonials, exaggerated product results, real-looking places that do not exist, deepfakes, copyrighted work used without permission, anyone’s likeness used without consent | Not allowed at all. A label does not make it acceptable. We will not produce it. |
| Label required | AI presenters and virtual influencers; a real person’s face or voice recreated with their consent; AI-made scenes or situations that shape how a product’s performance is understood; a product shown before it exists; sound effects central to what the product does | Allowed, and must be labelled. We tell you which, and supply the wording. |
| No label needed | Routine editing and colour correction, decorative backgrounds, ambient music, obviously fantastical effects, AI-assisted copywriting, and accessibility uses such as subtitles, captions and translations | No label needed. |
Most of what we make — AI UGC ads, AI product videos, AI commercials — sits in the "label required" category. So most of our deliveries come with a label recommendation.
What we will not produce
We will decline these, and we would rather say so upfront than halfway through a project. Most are prohibited by ASCI; one is our own standard.
- A fabricated testimonial. ASCI prohibits making it look as if a real person used, experienced, approved or recommended a product when they have not. A realistic AI presenter saying "I used this and it worked for me" does exactly that, because viewers take them for a real customer. A label does not fix it.
- Results a product does not actually deliver — for example, an AI before-and-after image showing a result the product has not been shown to achieve. No disclaimer fixes it.
- A real person’s face or voice without their written consent. This includes celebrities, public figures, your own staff, and your customers.
- Places or facilities that do not exist, presented as real.
- Our own standard: an AI-generated doctor, dentist, dietician or other professional endorsing a product. It implies expertise that nobody actually has.
How we label
Where a label is needed, we supply the wording with your delivery. ASCI accepts the labels ad platforms provide, or your own wording, as long as it accurately tells the viewer. The standard forms are:
- "Video created using AI" — where the scene, presenter or performance is generated.
- "Video enhanced using AI" — where real footage has been materially altered.
- "Sponsored by [brand]" — additionally required where the content is a paid product suggestion.
The label has to be genuinely visible to a viewer, not buried in small text at the end. For audio-only content, the disclosure is spoken. We build it into the deliverable where you want it burned in, or supply it as a separate overlay if you would rather control placement.
Provenance metadata
AI generation tools embed metadata marking a file as synthetically generated. The 2026 rules require that this is not stripped out. We do not remove it, and we ask that you do not either. Some editing and compression workflows discard metadata by accident, so if you re-encode our files, check that it survived.
Who declares it when the ad goes live
Meta, YouTube, Instagram and TikTok each have their own setting for declaring AI-generated content at upload. That declaration is separate from the on-screen label.
Unless we are managing publishing for you, ticking that box is your responsibility — you control the ad account. We always tell you whether your video needs it, so you are never guessing.
Using a real person
We can build an AI presenter based on a real person, such as your founder, a staff member or a customer, but only with that person’s written consent, given in advance.
We provide a short consent form for this. It records who the person is, what their likeness or voice will be used for, where it can appear, and for how long. This protects the person, you and us. Without a signed form, we will not proceed.
Content using a real person’s likeness always carries a label, even with consent.
Consent is a legal requirement here, not just a courtesy. Indian courts have repeatedly granted injunctions in 2026 against the unauthorised use of a person’s name, image and voice, including through AI-generated deepfakes and voice cloning, under what are known as personality rights. This applies to private individuals as well as public figures.
The tools we use
We use commercial AI generation tools for video, voice and image work, alongside conventional editing software. Our practice is to use a tool for client work only on a plan whose terms permit commercial use of the output, and to stop using it for client work if its terms change so that this is no longer true.
Which specific tool is used depends on what a shot needs. Your brief and materials are submitted to these providers to make your video. Each provider may collect and use what is submitted under its own terms and privacy policy, which we do not control, so we cannot guarantee the privacy of material once it reaches an AI tool. See our Privacy Policy. If you have a policy about particular tools or countries, tell us before we start and we will work within it.
What you are responsible for
- Product claims. You tell us what your product does; you must be able to substantiate it. We make the video, you stand behind the claim.
- Consent for any real person featured, unless you have asked us to obtain it.
- Declaring AI content at upload, unless we are managing publishing.
- Following the advertising rules of the markets you run in, which may differ from India’s.
Why we publish this
Most agencies have not written this page. The rules are recent, and compliance is easy to ignore until something goes wrong. The ASCI Code applies to everyone involved in making an ad, agencies included — so this is not a risk we can hand to you, and not one you should want to carry alone.
Being clear about it upfront is cheaper for both of us than finding out later.
Questions
If you are unsure whether something you want is allowed, ask before you commission it. Email support@naziadigital.com and we will tell you straight, including if the answer is that we cannot make it.