After Varun Dhawan: the Delhi High Court's deepfake personality-rights doctrine consolidates across the 2026 cluster

The Akkineni Naga Chaitanya order of 29 May 2026 — issued by Justice Jyoti Singh of the Delhi High Court — is the latest entry in what has become, over the first half of 2026, a tightly-packed line of personality-rights dispositions confronting the specific challenges of AI-generated content. Read in isolation, it is an interim summons order with an indicated injunction architecture. Read alongside the Swami Ramdev disposition of February, the Shashi Tharoor takedown direction of 10 May 2026, the protective orders in respect of Sunil Gavaskar and Arjun Kapoor, and the Varun Dhawan disposition of late May, it forms a doctrinal cluster whose contours are now visible to practitioners.

The Naga Chaitanya order

The plaintiff — Akkineni Naga Chaitanya, a Telugu film actor — moved the Delhi High Court against a combination of identified and unidentified defendants. The grievance categories cluster around contemporary patterns of identity misappropriation: pornographic domain names, webpages and metadata using the actor's identity; fabricated audio-visual content produced through AI tools and deepfake technology that, on the plaint's case, depicted the actor in a false, defamatory and misleading manner; voice-cloning content; and defamatory videos circulating on YouTube.

Justice Singh indicated that the court would pass an interim order protecting the plaintiff's personality rights. The next hearing was set for 30 September 2026. The order's most significant operational element is the Court's contemplation of a dynamic injunction — an architecture under which the plaintiff can move against future violations of the injunction without re-filing a fresh suit each time new infringing material surfaces.

The dynamic-injunction device is not new in Indian intellectual-property jurisprudence — it has been developed across copyright takedown matters, particularly against rogue websites — but its migration into personality-rights matters is doctrinally significant. The harm pattern with deepfakes is iterative: a fresh URL, a fresh upload, a fresh platform. A traditional injunction, requiring a fresh application each time, is structurally ill-suited to the speed of the harm. The dynamic injunction supplies a procedural pathway that matches the substantive problem.

The bench's posture, as reported in the court-reporting press, was that public life does carry vulnerability, but the protective architecture has its limits. That posture is doctrinally important. It contains, in compressed form, the limiting principle that the Delhi High Court has begun to articulate alongside the substantive protection. Public life carries a measure of consented exposure; not every reference to a public figure is actionable; the harm against which the Court will protect must be calibrated against the public's interest in commenting on, satirising, and engaging with public figures.

The cluster: a brief tour of the 2026 line

The Naga Chaitanya disposition is most useful read against the rest of the 2026 cluster.

Shashi Tharoor (10 May 2026). The Delhi High Court directed X — formerly Twitter — to take down an AI-generated deepfake video purportedly showing the Congress MP Shashi Tharoor praising Pakistan's diplomacy. The Court additionally directed Meta to ensure that certain offending reels on Instagram remained inaccessible. The harm pattern — fabricated political speech attributed to a public political figure — illustrates the doctrine's reach beyond the entertainment industry.

Sunil Gavaskar. In a separate matter, the Delhi High Court protected the cricketer's personality rights against unauthorised AI-deepfake use across social media and e-commerce platforms. The substantive grievance — commercial deployment of a deepfake likeness — fits the doctrine's commercial-gain limb.

Arjun Kapoor. A further disposition extended personality-rights protection against deepfake exploitation involving the actor's likeness.

Swami Ramdev (February 2026). Earlier in the year, the Delhi High Court restrained unauthorised AI-deepfake use of the yoga-teacher-cum-public-figure's persona. The order operated against a familiar pattern of misappropriation for commercial and ideological purposes.

Varun Dhawan (late May 2026). Justice Jyoti Singh — the same judge before whom Naga Chaitanya would shortly come — had indicated an interim order protecting the actor's personality rights against unauthorised merchandise, AI-generated content, deepfakes, pornographic content and morphed images. The Varun Dhawan disposition is part of the immediate doctrinal background to Naga Chaitanya.

Read together, the cluster discloses a remarkable doctrinal consistency. The same protected attributes recur — name, image, likeness, voice, mannerisms — and the same triad of actionable harms recurs: defamatory content, sexually explicit content, unauthorised commercial deployment.

The constitutional and statutory architecture

The doctrinal architecture the Delhi High Court has built rests on two pillars.

The constitutional pillar locates personality rights in Article 21 — the right to life and personal liberty, read to include the right to reputation, dignity, and the protection of identity attributes — and in Article 19 — the right to freedom of expression, read to include the right to control the commercial exploitation of one's likeness, and against which competing free-speech interests of the public must be balanced.

The statutory pillar engages the Information Technology Act, 2000 and the rules made under it, which now supply the takedown architecture for intermediaries; the Copyright Act, 1957, where the misappropriated content infringes a registered work; and the Trade Marks Act, 1999, where the public figure's name has been registered or is recognisable to the public as a source identifier.

The cluster does not, in any single disposition, exhaustively map this architecture. But across the line of cases, the components recur, and a practitioner approaching the doctrine from outside can read its constituent elements.

The limiting principle

The single most important doctrinal development in the cluster is the Court's articulation of the limiting principle. The bench's posture in Naga Chaitanya — that public life carries vulnerability, but the protection has its limits — captures the doctrinal stance across the 2026 line.

The limiting principle has three components, which can be inferred from the dispositions:

• The harm must be of an actionable kind: defamatory, sexually explicit, or for unauthorised commercial gain. Not all uses of a public figure's likeness fall within these categories.
• The protected interest is identity, not reputation in general: the doctrine does not displace defamation law; it operates alongside it where the harm is identity-misappropriating.
• The injunction architecture is calibrated to the harm: a dynamic injunction in deepfake matters, a takedown direction in defamatory-content matters, a commercial-injunction in merchandising matters.

The principle is doctrinally important because it prevents the personality-rights doctrine from collapsing into a general anti-speech weapon. The Court has been careful, across the cluster, to maintain the distinction between commercial misappropriation and legitimate public discourse.

The procedural device: how dynamic injunctions work

For the practitioner, the dynamic-injunction architecture is the single most operationally significant element. A dynamic injunction allows the rights-holder, after obtaining the initial order, to take fresh URLs, platforms, or instances of infringing content to a court-supervised process — typically through the intermediaries directly, with the Court's order serving as the authority — without filing a fresh suit.

The architecture's appeal in deepfake matters is that the harm is fundamentally iterative. A defendant who is restrained from one URL will surface on another within hours; the traditional injunction model, requiring a fresh interim application each time, is too slow. The dynamic injunction shifts the architecture from one of repeated litigation to one of structured takedown.

The Court's increasing comfort with the device — across copyright, now into personality rights — should be read alongside the Information Technology framework's takedown architecture, under which intermediaries are obligated to act on judicial orders.

What practitioners should do

For a client who is a public figure — actor, politician, sportsperson, ideological leader — the operational implications are now clear.

First, file early. The cluster shows that the Delhi High Court is receptive at the first indication of deepfake misuse; the harm pattern is such that delay materially worsens the position.

Second, frame the relief in dynamic-injunction terms. Identify the platforms — X, Meta, YouTube, the relevant e-commerce platforms — by name in the petition. The Court's order will, on the Naga Chaitanya template, contemplate intermediary compliance directly.

Third, attach the substantive harm to one of the three actionable limbs: defamation, sexual content, or commercial misappropriation. The Naga Chaitanya limiting principle suggests the Court will not extend the doctrine to harms that fall outside these categories.

Fourth, build the evidentiary record. Deepfake content is, by its nature, both transient and reproducible. Preserving the URLs, the timestamps, and where possible the underlying file artefacts is essential to the relief stage.

What the cluster does not decide

Three questions remain open across the cluster.

The first is the post-mortem question — whether personality rights survive the rights-holder's death, and if so for how long. None of the 2026 cluster engages this question; it will, at some point, come up.

The second is the satire and parody question. The cluster's limiting principle suggests the Court is alive to it, but no disposition in the cluster has had to draw the line between protected satire and actionable misappropriation in a case where the content was unambiguously satirical.

The third is the inter-platform enforcement question. The Court has, across the cluster, named X, Meta, YouTube and others. The compliance architecture across jurisdictions — particularly where the platform's holding entity is not within India — is an enforcement question that has not yet been comprehensively litigated.

These will come. The cluster is the doctrinal foundation on which the answers will be built.

Related editorial pieces

• Personality rights in the digital age: Delhi High Court protects Varun Dhawan against AI-generated misuse