• Download the draft guidelines here. 

The Advertising Standards Council of India (ASCI) has released draft guidelines for the responsible labelling of AI-generated content in advertising, proposing a risk-based framework that determines when disclosures are necessary. The industry body said the guidelines aim to ensure advertisements remain honest, decent, safe, and fair as brands increasingly deploy AI tools, while avoiding “consumer label fatigue”. ASCI has proposed that certain high-risk advertisements remain prohibited regardless of whether they carry an AI disclosure label.

Prohibited AI use:

• Fabricated endorsements or testimonials.
• Exaggerated product results or features through claims or visual representations that create a misleading impression.
• Fake but realistic-looking locations, such as a hotel or facility that does not exist.
• Deepfakes, copyrighted content or a person’s likeness used without consent.
• AI-generated fictional authority figures with identifiable cues, such as a fake doctor endorsing a supplement and implying medical expertise.

Where AI labelling is required: 

• Virtual or AI-generated influencers and brand ambassadors.
• Replication of a real person’s likeness or voice, even where consent has been obtained.
• AI-generated visuals used to demonstrate product performance must accurately reflect real-world performance.
• Entirely AI-created events, settings or situations presented realistically.
• Demonstrations of products that do not yet exist, such as an unbuilt housing project.
• AI-generated exaggerated sound effects that are highly relevant to a product’s core features, for example, in an ad for headphones.
• Paid or sponsored AI-generated product recommendations, which must be labelled as “sponsored by”.

Where AI labelling is not required:

• Routine edits such as colour correction, noise reduction, standard blemish removal and minor lighting adjustments.
• Decorative AI-generated backgrounds, abstract skylines, ambient music, jingles and unrelated background sound effects.
• Clearly fantastical or unrealistic elements, including dragons, fairies and other obviously fictional effects.
• Administrative and accessibility uses carried out in good faith, including advertising copy generation, audio descriptions for visually impaired users, and document preparation.

What MeitY’s draft synthetic media rules require

• The Ministry of Electronics and Information Technology’s (MeitY) February 2026 amendments brought synthetically generated information (SGI), including deepfakes, within the IT Rules’ due diligence framework, explicitly treating synthetic media as a form of online information subject to existing intermediary obligations.
• The rules significantly tighten compliance timelines, reducing content takedown timelines from 36 hours to three hours, grievance resolution timelines from 15 days to seven days, and certain emergency response timelines from 24 hours to two hours and from 72 hours to 36 hours.
• Intermediaries that enable the creation or sharing of SGI must deploy reasonable technical measures, including automated tools, to prevent unlawful synthetic content.
• Platforms must act against synthetic media containing child sexual abuse material, non-consensual intimate imagery, false documents, deceptive deepfakes, privacy violations, or content relating to explosives, arms or ammunition.
• For permitted AI-generated content, platforms must prominently label the content, provide audio disclosures where applicable, and embed provenance information, metadata and unique identifiers wherever technically feasible.
• Intermediaries must not permit the modification, suppression or removal of AI labels and provenance markers.
• Once a platform becomes aware of a violation, whether through user complaints, notices or its own detection mechanisms, it must take expeditious action, including disabling access, suspending accounts, preserving evidence and reporting violations where required.
• Significant social media intermediaries must require users to declare whether content is synthetic, verify those declarations using technical measures, and ensure synthetic content carries clear labels before publication.
• Platforms must issue enhanced user notices every three months, warning users about content removal, account suspension, identity disclosure and potential legal consequences for violations involving synthetic media.
• However, the rules exempt routine editing, colour correction, accessibility tools, document preparation, educational materials and other good-faith uses that do not materially alter or misrepresent underlying content.

Why does this matter? The draft guidelines arrive as AI-generated advertising and deepfake scams become increasingly difficult for consumers to recognise. McAfee found that 90% of Indians have encountered fake or AI-generated celebrity endorsements, with victims losing an average of Rs 34,500 to such scams. Meanwhile, Brazilian authorities recently uncovered a fraud operation that allegedly earned millions of dollars through Instagram advertisements featuring deepfakes of supermodel Gisele Bündchen and other celebrities. ASCI’s proposed framework, therefore, seeks to draw a clearer line between acceptable AI-assisted advertising and deceptive synthetic content. It also aims to ensure that disclosures do not become so common that consumers simply stop paying attention to them.

Also read:

• On Advertising Inside AI
• How will Mandatory Labels on Synthetic Content Affect Digital Ad Campaigns? #NAMA
• Google demo shows experimental AI ad rendering onto videos, other AI capabilities in ad tech