Specialists say the important thing modifications will impose stricter due diligence obligations on intermediaries. File | Picture Credit score: Getty Pictures
The Ministry of Info Expertise on Wednesday (October 22, 2025) proposed new guidelines to label, monitor and import deepfake AI-generated content material underneath IT guidelines, a transparent step in the direction of making certain the authenticity of digital content material, cybersecurity and expertise consultants stated.
The division’s proposed guidelines would require social media firms to require customers to declare whether or not they’re importing deepfake content material created by synthetic intelligence (AI) that might hurt people, organizations, or governments.
Supreme Courtroom Justice Pavan Duggal, who makes a speciality of cyber legislation, cyber crime legislation, cybersecurity legislation, and AI legislation, stated: hinduism: “The proposed Indian Info Expertise Guidelines, 2021 marks a historic leap ahead in digital laws in our nation, lastly confronting the complicated menace of deepfakes and artificial content material.”
He stated that for the primary time, the draft amendments to India’s cyber legal guidelines acknowledge and clearly outline “synthetically generated data” as computer-altered content material disguised as real. This can be a long-awaited shift that aligns the legislation with digital actuality.
“In an AI-driven period the place it’s nearly indistinguishable between fabricated and actual information, this definitional certainty is transformative. This modification captures the pressing have to crack down on artificial content material that, when misused, can destabilize belief, unfold misinformation, and undermine digital integrity,” Duggal stated.
Crucially, he says, these modifications will impose stricter due diligence obligations on intermediaries. Platforms that facilitate the creation of artificial content material should be sure that all such elements are labeled unambiguously and embedded with persistent metadata or identifiers. “This labeling, which covers at the least 10 p.c of visible and audio interfaces, isn’t just a formality, however a decisive step to make sure public consciousness,” Duggal asserted.
“Labeling AI-generated supplies and embedding non-removable identifiers will assist customers distinguish between actual and artificial content material,” stated Mahesh Makhija, Companion and Expertise Consulting Chief, EY India.
He additional stated that it will lay the inspiration for accountable AI adoption and that these measures will give companies the arrogance to innovate and scale AI responsibly.
“The subsequent step is to determine clear implementation requirements and collaboration frameworks between authorities and business to make sure the foundations are sensible, scalable, and supportive of India’s AI management ambitions,” Makhija added.
Echoing related sentiments, Duggal additionally weighed in. Nevertheless, implementation requires technical sophistication, cross-platform requirements, sturdy enforcement, and worldwide coordination.
Duggal additional stated that the brand new accountability framework will additional strengthen the system. “If an middleman deliberately permits or ignores unmarked artificial content material, it is going to be deemed to have didn’t train due diligence, placing the vital Part 79 protected harbor immunity in danger.”
Underneath the proposed regime, social media platforms will now have three obligations: Customers should disclose artificial origin. Platforms ought to implement technical instruments to confirm such disclosures. Each time artificial origin is confirmed, a transparent and conspicuous artificial label have to be displayed.
