Fair Dealing for LLMs and a Ghost Regulator: How the Courts are Rewriting India's Tech and IP Landscape in 2026

The Collision of Constitutional Privacy and Copyright in the AI Era

If you are advising tech clients, media houses, or data fiduciaries in 2026, the ground beneath your feet is shifting rapidly. The Supreme Court and the High Courts are currently untangling two of the most consequential legal knots of our generation: the constitutional validity of the Digital Personal Data Protection (DPDP) Act’s intersection with the RTI Act, and the existential threat generative AI poses to the Copyright Act, 1957.

The headline for practitioners is twofold: First, the Supreme Court has unequivocally refused to stay the operation of the DPDP Act and the 2025 Rules while referring challenges to a larger bench. Second, the Delhi High Court has just fired the starting gun on AI copyright litigation, handing a massive prima facie victory to LLM developers under the guise of "fair dealing."

The DPDP Act vs. RTI Act: A Constitutional Showdown

The Supreme Court’s decision to refer the DPDP Act challenges to a larger bench is not just procedural housekeeping—it is an acknowledgment of a fundamental statutory clash. At the heart of this litigation is the DPDP Act’s amendment to Section 8(1)(j) of the Right to Information (RTI) Act, 2005.

Historically, the RTI Act allowed the disclosure of personal information if a larger public interest justified it. The DPDP Act obliterated that nuance, creating a blanket exemption for "personal data." Petitioners are rightly arguing that this fundamentally alters India's public-interest disclosure framework. Furthermore, a fresh challenge to Section 44(3) of the DPDP Act has prompted the Supreme Court to issue notice regarding the masking or deletion of currently available public data.

The Supreme Court has now framed the ultimate question: What exactly constitutes “public data” versus “personal data” under this new regime?

"We are witnessing a dangerous regulatory vacuum. The law is operational, compliance mandates are live, yet the Data Protection Board (DPB) remains a ghost town with no appointed Chairperson or Members."

This operational absurdity was recently highlighted in the Kerala High Court, which demanded to know the status of the DPB in a dispute over Digi Yatra passenger data. For practicing lawyers, this means you must advise your Data Fiduciary clients to strictly comply with the 2025 Rules—investing heavily in the booming DPDP compliance tooling market—even while the adjudicatory body meant to enforce these rules does not practically exist. The Supreme Court's refusal to grant a stay means there is no grace period. If your clients are waiting for the constitutional challenge to play out before overhauling their consent managers, they are walking into a trap.

Opt-Outs and Edu-Tech: The APAAR Mandate

The DPDP Act's teeth are already showing in specific sectors. Take the Supreme Court's recent directions regarding the APAAR (Automated Permanent Academic Account Registry) ID for students. The Court explicitly brought student data collection under the rigorous strictures of the DPDP Act, mandating opt-out language in parental consent forms.

For lawyers representing Edu-tech companies or educational institutions, this is a clear directive: implicit consent or bundled "take-it-or-leave-it" terms are dead. Consent architectures must be unbundled, verifiable, and freely revocable, shifting the market standard from mere consent management to provable compliance evidence.

Delhi High Court on AI: Stretching Section 52(1)(a)

While the Supreme Court wrestles with privacy, the Delhi High Court has delivered a seismic shock to intellectual property practice. In the ongoing copyright battle involving ANI and OpenAI (ChatGPT), the Court held, prima facie, that the ingestion and storage of copyrighted works for training Large Language Models (LLMs) falls under Section 52(1)(a) of the Copyright Act, 1957.

Let us be blunt: categorizing the mass scraping of proprietary news content by a multi-billion-dollar commercial AI entity as "fair dealing for the purpose of private use, including research" is a staggering doctrinal stretch. Section 52(1)(a) was drafted for scholars and critics, not algorithmic data-mining at a global scale.

However, from a pragmatic standpoint, this is a highly calculated move by the judiciary. By adopting a permissive interpretation of fair dealing, the Delhi High Court is ensuring that India does not become a hostile jurisdiction for AI innovation.

What this means for IP practitioners:

If you represent content creators, publishers, or media houses, traditional infringement notices against AI scrapers are currently fighting an uphill battle against this prima facie view. You must immediately pivot your strategy. Instead of relying purely on ex post copyright enforcement, advise clients to adopt robust technological protection measures (TPMs), aggressive payload encryption, and restrictive API licensing agreements. Contract law and technological barriers will be your best defense where copyright doctrine is currently failing you.

The Verdict for Practitioners

2026 is shaping up to be the year where statutory text meets operational reality. The Supreme Court's framing of "public vs. personal" data will dictate the future of investigative journalism and corporate transparency in India. Meanwhile, the Delhi High Court's AI ruling forces a complete rewrite of IP monetization strategies.

Lawyers must stop waiting for the dust to settle. Draft your privacy policies assuming the DPDP Act will survive constitutional scrutiny intact, but structure your IP licenses assuming the Copyright Act will offer little protection against the AI machine.