Draft ASCI guidelines for SGC in advertising

India does not have a standalone law to regulate AI. However, this does not mean that the space is unregulated. To regulate synthetically generated information (SGI), the Ministry of Electronics and Information Technology has recently amended the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 (IT Rules).

Further, for AI use specifically in advertisements, the Advertising Standards Council of India (ASCI), a self-regulatory body, has subsequently released Draft Guidelines for Responsible Labelling of Synthetically Generated Content in Advertising (draft guidelines) for public consultation. The draft guidelines align with amendments in the IT Rules.

The draft guidelines regulate synthetically generated content (SGC), which includes artificially created, modified, or materially altered content, including deepfakes, AI voices, etc.

Three-tier risk rules for ads

Further, the draft guidelines propose a three-tier risk framework:

1. 1. High-risk advertisements (e.g. fabricated endorsements and unauthorised deepfakes) are prohibited even with AI labels.
   2. Medium-risk advertisements (e.g. virtual influencers, AI-generated likeness or voices, etc.) must be labelled as per the ASCI guidelines for disclaimers.
   3. Low-risk advertisements (e.g. minor enhancements) do not require labelling.

Any non-compliance would be actionable under the ASCI Code for Self-Regulation of Advertising Content (ASCI code).

ASCI compliance becomes business essential

While the ASCI code and guidelines do not have the force of law, the ASCI enjoys substantial industry acceptance and has consistently adopted an active enforcement approach. Consequently, compliance with ASCI standards in the case of advertisements is often considered essential.

Labelling ads is not new, as some companies like Meta already have a policy for labelling such content. However, in terms of their impact, the draft guidelines, if adopted in their current form, would push businesses to implement mechanisms to review content and contractually ensure with advertising agencies, influencers, endorsers, celebrities and other marketing partners that responsibility for compliance, disclosure, and associated liabilities is appropriately allocated.

The draft guidelines adopt a risk-based approach to labelling content, placing the compliance burden on brands/advertisers, and requiring them to exercise a greater degree of judgement while determining if their content triggers a disclosure obligation.

Draft AI rules widen accountability

Similar to the IT rules, the draft guidelines recognise that minor enhancements and good-faith editing do not ordinarily pose a material risk. However, these frameworks differ in their conceptual approach. Under the IT rules, such modifications fall outside the definition of SGI altogether. However, the draft guidelines define SGC broadly to encompass such modifications but classify them as low-risk content that does not require labelling. While the practical outcome may be similar, advertisers would have to be mindful of the rationale underlying such decisions.

Further, specific consent would have to be obtained for using someone’s likeness. Moreover, compliance with the draft guidelines, if adopted, is unlikely to be satisfied merely by affixing a disclosure label, as the ASCI may assess the substantive content of an advertisement irrespective of the disclosure/label. Accordingly, advertisers/brands will need to ensure that the use of AI-generated content does not result in misleading claims.

The position is particularly significant when viewed alongside the broader consumer protection framework. The Central Consumer Protection Authority (CCPA) has requested ASCI to forward any advertisement that does not comply with the ASCI Code and could potentially violate the Consumer Protection Act, 2019, for regulatory action. While there is presently no indication that the CCPA intends to issue a dedicated framework governing AI-generated content, the draft guidelines may prompt greater regulatory attention from the CCPA, which may issue broader guidelines on the use of AI.

Nevertheless, the draft guidelines indicate a shift towards accountability while using AI-generated content. Accordingly, brands should approach compliance from a risk management perspective rather than a mere labelling requirement.

Rishikaa is a senior associate and Anindita Deb is an associate at ADP Law Offices