IAMAI Raises Red Flags On MeitY’s Draft AI Labelling Rules

Shortly after the government issued its advisory on mandatory AI labelling to curb AI-generated deepfakes, industry body IAMAI has raised several concerns about the draft amendments to the IT Rules circulated late October.

In a note accessed by Inc42 from industry sources, the association said that existing provisions under the IT Act and IT Rules already address unlawful synthetic content, and cautioned that introducing an additional synthetic-content framework is “vague, expansive and difficult to implement at scale”.

In its submission, IAMAI noted that the proposed definition of synthetic and manipulated content (SGI) is so broad that it could include routine digital edits made for accessibility, moderation or basic user adjustments. It added that mandatory watermarking, metadata insertion and verification rules could degrade user experience, lead to privacy issues and create heavy compliance burdens, especially for startups.

“We respectfully submit that the inclusion of Rule 2(1A) is unnecessary and may introduce widespread ambiguity and greater compliance burdens, despite the existing due diligence framework already equipping intermediaries – particularly social media platforms – to address SGI effectively,” the association said in its filing.

IAMAI pointed out that Section 66D of the IT Act, which addresses impersonation using computer resources, and Section 79, which governs safe harbour and takedown of unlawful content, already cover cases of harmful deepfakes, making additional prescriptive mechanisms unnecessary.

Under the draft proposal, the MeitY has sought visible or audible labels on at least 10% of synthetic content, user declarations, and expanded due-diligence requirements for intermediaries. IAMAI argued that such obligations are technically infeasible due to the lack of industry standards and the immaturity of watermarking, metadata and AI-detection technologies.

The industry body also said the advisory fails to distinguish between platforms hosting third-party AI-generated material and companies offering first-party AI services. It warned that conflating both categories expands the definition of an intermediary and unnecessarily subjects first-party AI providers to safe-harbour-linked obligations.

Another concern is the broad scope of the SGI definition, which IAMAI said could sweep up non-deceptive editorial changes such as colour correction, cropping or speech-to-text accessibility edits.

Digital rights group IFF has echoed similar concerns in recent weeks, warning that mandatory labelling may lead to compelled speech and de-facto general monitoring obligations.

In its recommendations, IAMAI urged the government to adopt tech-neutral definitions, remove SGI-specific clauses and mandatory labelling requirements, and avoid pre-emptive content verification until global technical standards emerge. At present, only the European Union and the US state of California mandate synthetic-content labelling.

The debate comes amid a surge in AI misuse in India, with deepfake-driven hoaxes, scams and reputational risks affecting public figures and ordinary users alike. Celebrities such as Aishwarya Rai Bachchan and Hrithik Roshan have already approached courts seeking protection of their publicity and personality rights, with interim relief granted in multiple cases.