Key facts
- ASCI lists deepfakes and a person's likeness without consent among prohibited uses, which stay prohibited even if an AI label is used1.
- ASCI also requires a label when a real person's likeness or voice is replicated, even with their consent, for personalised messaging1.
- Google Ads treats business founders and CEOs as public figures, asks for documented consent before their name, image, video or voice is used to suggest an endorsement, and can suspend accounts without prior warning for misrepresentation2.
- In an interim order dated 20 February 2026, the Delhi High Court barred use of Kajol's name, image, voice and likeness, including through deepfakes, AI-generated content and chatbots, and gave platforms 72 hours to remove flagged content3.
- India's IT Rules require tool providers to use technical measures against synthetic content that falsely depicts a person's identity, voice, conduct or statements in a way likely to deceive4.
Can I put myself or my founder in an AI ad?
Yes, with that person's real agreement, and you should have it in writing. Google's ad policy lists business founders and CEOs among public figures and asks advertisers to hold documented consent they can prove before using a name, image, video or voice to suggest an endorsement2. A founder who agrees today can withdraw tomorrow, so the consent letter should say what the AI version may say and do, where it will run, for how long, and how the person can ask for it to come down.
A label may still be needed. ASCI requires one when a real person's face or voice is replicated, even with consent, for personalised messaging1. On YouTube, making a real person appear to say or do something they did not must be disclosed for uploaded videos5. Meta shows an AI info label next to Sponsored on ads with an AI-generated photorealistic human made with its tools6.
The words matter as much as the face. ASCI prohibits using AI to create or attribute an endorsement that falsely says a real person used, approved or recommended a product1. Do not script claims the person has not made.
Can I use a celebrity or a lookalike?
Not without their consent. ASCI prohibits deepfakes and likeness without consent whether or not a label is used1. Google's policy gives AI-generated videos of popular YouTubers as an example of a misleading endorsement2. A lookalike face or a soundalike voice that makes viewers think the real person endorsed your product meets the same objection: ASCI bars fabricated endorsements and Google targets any suggestion of endorsement.
The Delhi High Court has been active. These are interim orders reported by news outlets, not final judgments, and we have not read the orders themselves:
- Kajol, 20 February 2026: Justice Jyoti Singh barred use of her name, image, voice and likeness, including through deepfakes, AI-generated content and chatbots, with 72 hours for platforms to remove flagged content3.
- Baba Ramdev, February 2026: an interim order protecting his personality rights, covering AI-generated content and deepfakes, with false endorsements among the concerns and Google, Meta and X told to remove listed links within 72 hours7.
- Aishwarya Rai Bachchan, 9 September 2025: an interim injunction against exploiting her name and likeness through AI, deepfakes or face morphing8.
- Ravi Kishan, July 2026: an ex parte ad interim injunction against unauthorised use of his name, image and likeness through AI and deepfakes9.
What about customers, employees or any other real person?
The same logic applies: get explicit consent before you replicate anyone. HeyGen, an AI avatar vendor, requires the explicit consent of the individual being represented, lets that person ask for removal of their likeness at any time, and makes users responsible for honouring such requests10. That is a vendor's rule, but it is a sensible baseline for any tool.
An AI-generated person is not a customer. Do not present a generated face as someone who bought your product; ASCI prohibits fabricated testimonials even with a label1.
Is there a law for personality rights in India?
Reports describe these rights as judicially recognised rather than set out in one statute, and LiveLaw's report of the Ravi Kishan order cites no statute9. That means outcomes depend on the facts and the court. The IT Rules add a platform-side layer: tools must use technical measures against synthetic content that falsely depicts a natural person's identity, voice, conduct or statements in a way likely to deceive4. If you are unsure, get a lawyer's view before you spend on production. This is general information, not legal advice.
What should I have in place before I publish?
A short pre-publish check for any ad with a real person's face or voice:
- Written consent that names AI use, the media, the territories and the duration.
- The script the person has approved, word for word.
- A label where the ad replicates a real person, following ASCI's examples1.
- A record of the tool, the source images and the date.
- A named contact and a process to take the ad down quickly on request.
- No testimonial, endorsement or before-and-after result that did not happen.
Our side, stated plainly
What KanavuLab does
Generated people and scenes in our studio work are disclosed as generated and are never presented as real customers. The storyboard is approved before anything is rendered, and you approve everything that goes out under your name.
We are not a law firm, so have your lawyer review any consent letter. Video work is scoped and quoted after a 20-minute call.
Follow-up questions
Can I make an AI version of our founder for ads?
Yes with the founder's documented consent. Google Ads asks for provable consent from public figures, and founders and CEOs are on its list2. ASCI may also require a label when a real person's likeness or voice is replicated1. This is general information, not legal advice.
Is an AI label enough to use a celebrity's face?
No. ASCI prohibits using a person's likeness without consent even if an AI label is used1.
Sources
Every source was opened before this page was published. Figures from vendors who sell in this market are labelled as such in the text.
- ASCI Guidelines for Responsible Labelling of Synthetically Generated Content in Advertising, 17 September 2026. Opened 12 October 2026
- Google Advertising Policies Help Unacceptable business practices: public figure impersonation. Opened 12 October 2026
- Storyboard18 Delhi High Court grants interim protection to Kajol's personality rights, orders takedown of AI deepfakes, 25 February 2026. Opened 12 October 2026
- MeitY, Government of India IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, updated 10 February 2026, 10 February 2026. Opened 12 October 2026
- YouTube Help Disclosing use of altered or synthetic content. Opened 12 October 2026
- Meta Help Centre How AI-generated images in ads are identified and labelled on Meta. Opened 12 October 2026
- Storyboard18 Delhi High Court protects Ramdev's personality rights, orders takedown of deepfakes and false endorsements, 25 February 2026. Opened 12 October 2026
- Storyboard18 Delhi HC backs Aishwarya Rai Bachchan, rules personality rights breach erodes celebrity dignity, 11 September 2025. Opened 12 October 2026
- LiveLaw Delhi High Court protects actor Ravi Kishan's personality rights, orders takedown of AI-generated obscene content, 7 July 2026. Opened 12 October 2026
- HeyGen (an AI avatar vendor) Moderation policy. Opened 12 October 2026
- MediaNews4U ASCI issues guidelines for labelling AI-generated content in advertising, 29 September 2026. Opened 12 October 2026