The Transparency Purge & The AI Loophole: The Supreme Court’s DPDP Stance and Delhi HC’s OpenAI Bombshell

The Collision of Privacy, Transparency, and Innovation

For Indian tech and media lawyers, 2026 is shaping up to be the year where the traditional silos of privacy, intellectual property, and constitutional law finally collapse into a single, chaotic domain. Two massive developments have just hit the docket: the Supreme Court’s refusal to stay the contentious Digital Personal Data Protection (DPDP) Act, 2023, and a startling observation by the Delhi High Court in the ongoing OpenAI v. ANI litigation regarding copyright and artificial intelligence.

Both developments revolve around a single, foundational question that will define Indian jurisprudence for the next decade: Who owns the data, and who gets to look at it?

The RTI Act’s Death by a Thousand Cuts

Let’s start with the Supreme Court. The apex court recently issued notice in multiple writ petitions—including those filed by the Mazdoor Kisan Shakti Sangathan and the Reporters Collective—challenging the DPDP Act and the recently notified DPDP Rules 2025. However, in a blow to transparency advocates, the Court categorically refused to grant an interim stay on the operation of the Act.

For practitioners, the focal point of this litigation is Section 44(3) of the DPDP Act. This provision is a Trojan horse that fundamentally alters Section 8(1)(j) of the Right to Information (RTI) Act, 2005.

Prior to this amendment, Section 8(1)(j) provided a nuanced balancing test: personal information could be exempted from RTI disclosure unless the Public Information Officer (PIO) was satisfied that the larger public interest justified its disclosure. It was a statutory embodiment of proportionality.

Section 44(3) of the DPDP Act wipes that balancing test off the books. It creates an absolute embargo on the disclosure of "personal information" under the RTI Act, effectively weaponizing the right to privacy to shield the administrative state from public scrutiny. By refusing to stay the masking or deletion of already-available public data, the Supreme Court is allowing a profound shift in the information ecosystem.

What this means for practice: If you represent journalists, NGOs, or corporate whistleblowers, your traditional RTI strategy is currently gutted. PIOs now have a blanket statutory excuse to reject any application that even tangentially touches upon "personal data"—a term defined so broadly under the DPDP Act that it can encompass almost any identifiable administrative record. The Supreme Court has indicated it will examine the boundary between public data and private data, but until a larger bench rules, expect a chilling effect on transparency litigation.

A Ghost Regulator and the November Compliance Mirage

While the State uses the DPDP Act as a shield, corporate India is left grappling with a phantom regulatory regime. The Data Protection Board of India (DPBI) has been legally established, yet as of mid-2026, it remains an empty shell with no appointed Chairperson or Members.

Despite this bureaucratic paralysis, the compliance clock is ticking. Phased implementation is underway, with strict consent-related compliance and Consent Manager registrations expected to go live in November 2026. Furthermore, the Supreme Court recently held that personal data handled under the government's APAAR Scheme remains firmly subject to the DPDP Act, prohibiting the disclosure of student information to private third parties without lawful basis.

The takeaway for corporate counsel: You cannot advise your clients to wait for the DPBI to wake up. The courts are already enforcing DPDP principles in real-time. Tech companies must overhaul their notice and consent architectures now, or risk massive statutory penalties once the Board is finally populated and looks to make an immediate regulatory splash.

OpenAI v. ANI: Is Scraping for LLMs Really "Fair Dealing"?

While the Supreme Court wrestles with the DPDP Act, the Delhi High Court has dropped a bombshell in the intellectual property space. In the high-stakes OpenAI v. ANI copyright infringement suit, the High Court observed that training a Large Language Model (LLM) on copyrighted literary works might fall within the ambit of Section 52(1)(a) of the Copyright Act, 1957.

Section 52(1)(a) exempts "fair dealing" with any work for the purposes of private or personal use, including research.

To put it bluntly: this is a staggering judicial stretch. Equating the ingestion of millions of copyrighted news articles by a multi-billion-dollar commercial entity to train a for-profit AI model with "private research" fundamentally misreads the legislative intent of the Copyright Act. Indian copyright law lacks the broad, flexible "fair use" doctrine of the United States; our "fair dealing" exceptions are exhaustive, specific, and have traditionally been interpreted narrowly.

Impact on IP Litigation: If this observation crystallizes into a binding precedent, it will gut the licensing market for Indian content creators and media houses. IP litigators representing publishers need to immediately pivot their strategy. You cannot rely solely on traditional reproduction rights under Section 14. You must forcefully argue that LLM training is a commercial exploitation that intrinsically alters the market for the original work, failing the fair dealing test under established precedents like Civic Chandran v. Ammini Amma.

The Bottom Line

We are witnessing the growing pains of India's digital jurisprudence. On one hand, privacy is being invoked via the DPDP Act to kill the RTI Act. On the other hand, traditional IP protections are being diluted to give generative AI a free pass under archaic "fair dealing" clauses.

For the sharp practitioner, the writing is on the wall. You can no longer practice IP, privacy, or constitutional law in isolation. The modern Indian lawyer must master the intersection of all three, because right now, that intersection is where the law is being rewritten.