'Deepfakes Disproportionately Target Women, Takedown Comes After Harm Is Done': Justice PS Narasimha

Speaking at an event, Supreme Court judge Justice PS Narasimha today raised a concern about the disproportionate targeting of women using deepfakes and lamented that harm has already occurred by the time takedown action is effected..

"Let's talk about the problem we are all facing, relating to deepfakes...non-consensual, synthetic imagery disproportionately targeting women, perpetuating gender-based violence and inflicting acute psychological harm. Our present legal framework responds reactively, but takedown comes after the harm has already occurred", the judge said.

Justice Narasimha was speaking at the 5th Soli Sorabjee Memorial Lecture organized by the Commonwealth Lawyers Association in collaboration with Nivaaran. The topic of the lecture was "Human Rights in the Empire of Artificial Intelligence".

Besides Justice Narasimha, the event was graced by Supreme Court judge Justice PB Varale, ASG Anil Kaushik, Senior Advocates Vikas Singh, Jaideep Gupta and Rajiv Dutta. It can be watched here.

At the outset of his address, Justice Narasimha reminisced his practice days and interactions with Soli Sorabjee. "One unique thing about Sorabjee is that all those who knew him always felt that he liked them very much. I was one of them. His underlying philosophy regarding the role of Attorney General was fundamentally rooted in institutional independence and constitutional morality...He approached the law not merely as a profession but as a public responsibility. His advocacy was marked by balance and restraint. Over a career of 7 decades, he was a fierce defender of civil liberties and human rights" the judge said.

Subsequently, he emphasized how artificial intelligence operates as a less visible form of power, affecting people's lives and breaching the concept of "informed consent". "AI confronts us with a more diffused and less visible form of power embedded in technological systems...empire of AI does not recognize national boundaries. AI does not displace power. It reorganizes itself in ways that are less visible but no less consequential", he said.

The judge also flagged the tendency of AI to result in "exclusion" if it is used by the State in distribution of welfare benefits. "As State turns to AI to improve efficiency in service delivery, [...] emerges in technological optimization and inclusive access. Institutions are relying on AI-driven platforms to identify beneficiaries, process claims and distribute entitlements...this has direct implication for human rights because welfare delivery is not simply an administrative function. It is central to realization of socio-economic rights. Application of AI can lead to exclusion."

The judge further stated that the guarantee of due process must be extended to algorithmic processes. He also referred to privacy concerns raised by use of AI. "Privacy can no longer be understood as mere secrecy or non-disclosure. It must be understood as control over personal data and the governance of the complex infrastructure that process information at large scale. DPDP Act does not fully grapple with deeper problems like inference."

The lecture ended with Justice Narasimha saying that institutions must ensure that integration of AI does not dilute their obligation to secure rights but instead reinforces principles of inclusion, fairness, accountability, that remain at the "core" of human rights.