The Privacy Shield or the Transparency Gag? How the DPDP Act is Rewriting RTI, While the Delhi HC Drops an AI Copyright Bombshell

The Collision of Privacy and Transparency

For practicing lawyers, 2026 is rapidly shaping up to be the year where theoretical technology law finally collides with the gritty reality of constitutional litigation. The Supreme Court’s recent moves regarding the Digital Personal Data Protection (DPDP) Act, 2023, and the Delhi High Court’s prima facie ruling on Generative AI, represent seismic shifts in how we practice law, gather evidence, and advise commercial clients.

The most pressing development for the Indian Bar is the Supreme Court’s February 2026 decision to refer a batch of petitions challenging the DPDP Act and the 2025 Rules to a larger bench. Crucially, the Court categorically refused to stay the operation of the law. While the challenge spans consent architecture and Data Protection Board enforcement, the true battleground—and the most immediate headache for litigators—is Section 44(3) of the DPDP Act.

Section 44(3): The Death Knell for Public Interest RTI?

To understand why Section 44(3) matters to your practice, we must look at the mechanics of the Right to Information (RTI) Act, 2005. Historically, Section 8(1)(j) of the RTI Act exempted "personal information" from disclosure unless the Central Public Information Officer (CPIO) was satisfied that the "larger public interest justifies the disclosure." It was a balancing test. It allowed investigative journalists, activists, and writ lawyers to access beneficiary lists, public servant asset declarations, and welfare scheme data to build robust PILs and expose corruption.

Section 44(3) of the DPDP Act surgically removed that public interest caveat. It amended Section 8(1)(j) of the RTI Act to create a blanket exemption for any "information which relates to personal information."

This is not merely a statutory tweak; it is the weaponization of privacy to shield state opacity. By refusing to stay this provision, the Supreme Court has temporarily handed public authorities an impenetrable shield.

In practice, CPIOs are already rejecting RTI applications en masse by citing the DPDP Act. The April 2026 PIL seeking interim relief against the masking and deletion of data already available on public portals highlights the panic on the ground. Welfare accountability systems—like NREGA muster rolls or PDS distribution lists—rely on public transparency. If you are a lawyer relying on RTI responses to annex critical fact-finding to your Article 226 petitions, your pipeline of evidence has just been severely choked. Until the larger bench resolves the distinction between "public data" and "private data" (an issue the SC framed in March 2026), expect the State to treat all data as strictly private.

Meaningful Consent is Not Just a Tick-Box

While the State uses the DPDP Act to block information, the Supreme Court has shown teeth when it comes to the State's own data collection. In the late July/early August 2026 ruling concerning the APAAR (Automated Permanent Academic Account Registry) Scheme, the Supreme Court delivered a crucial reality check on what constitutes "consent" under Section 6 of the DPDP Act.

The Court held that the collection and processing of student data by the State remains strictly subject to the DPDP Act. More importantly, it ruled that consent forms must provide parents or guardians a genuine, unpenalized option to withhold consent. This is a vital precedent. It signals to corporate Data Fiduciaries and State instrumentalities alike that the era of "take-it-or-leave-it" bundled consent is over. If you are drafting privacy policies or consent architectures for tech clients, the APAAR ruling dictates that consent must be affirmative, granular, and capable of being denied without the denial of the underlying service.

The AI Copyright Bombshell: ANI v. OpenAI

Parallel to the privacy debates, the Delhi High Court has fundamentally disrupted Intellectual Property practice. In the highly anticipated 2026 ANI v. OpenAI litigation, the High Court held, prima facie, that the ingestion of copyrighted material to train Large Language Models (LLMs) could fall within the ambit of Section 52(1)(a) of the Copyright Act, 1957.

Section 52(1)(a) provides that "fair dealing" with any work for the purposes of "private or personal use, including research" does not constitute copyright infringement.

Let us be direct: applying a "private or personal use" exception to a multi-billion dollar commercial AI model scraping the internet to build a competing commercial product is a massive, arguably strained, statutory leap. The Indian "fair dealing" doctrine is notoriously narrower than the American "fair use" doctrine. By shoehorning LLM training into "research," the Delhi High Court is actively choosing to prioritize technological innovation over traditional copyright maximalism.

What this means for IP Practitioners: If you represent publishers, media houses, or authors, the traditional cease-and-desist playbook is obsolete. You can no longer rely on a straightforward reproduction/infringement argument under Section 51. You must now be prepared to litigate the boundaries of "fair dealing" in the context of machine learning. Conversely, if you represent AI startups, this prima facie ruling is a golden shield, granting immense breathing room to innovate without the immediate threat of crippling infringement injunctions.

The Bottom Line

Indian technology law in 2026 is defined by courts attempting to retrofit 20th-century statutes (like the Copyright Act and the RTI Act) onto 21st-century paradigms (GenAI and Data Fiduciaries). The DPDP Act is proving to be a double-edged sword—imposing strict consent mandates on the private sector while simultaneously acting as a transparency gag for the government. Practitioners must adapt swiftly: diversify your fact-finding beyond the RTI Act, audit your clients' consent flows immediately, and prepare for a fundamentally altered landscape in copyright litigation.