This paper has been written by Atul Verma.
This paper has been selected for LLJ Publications.
Abstract
The Bharatiya Sakshya Adhiniyam, 2023 (“BSA”), which replaced the Indian Evidence Act, 1872 with effect from 1 July 2024, restructures the law governing electronic and digital evidence in India through Sections 2(1)(d), 57, 61, 62 and 63. This paper undertakes a doctrinal and analytical examination of these provisions, tracing their lineage from Sections 65A and 65B of the repealed Act and situating them within the trajectory of judicial interpretation that runs from State (NCT of Delhi) v. Navjot Sandhu through Anvar P.V. v. P.K. Basheer, Shafhi Mohammad v. State of Himachal Pradesh, and Arjun Panditrao Khotkar v. Kailash Kushanrao Gorantyal, culminating in the Supreme Court’s 2026 pronouncement in Pune Bar Association v. Union of India upholding the constitutionality of the dual-certification and hash-value regime under Section 63(4). The paper examines three interlocking themes: first, the statutory conditions for admissibility of computer output and the continuing debate on whether Section 63 is an exhaustive code or merely one permissible mode of proof, particularly in light of the overriding language of Section 61; second, the practical and forensic dimensions of authentication, including hash-value verification, chain of custody, and the unresolved question of who qualifies as a competent “expert” for Part B certification; and third, the emerging jurisprudential and regulatory challenges posed by deepfakes and artificial-intelligence-generated content, which strain a framework designed primarily for static computer output. Drawing on statutory text, case law, the 2024 Law Commission material, and post-2024 judicial and regulatory developments up to mid-2026, the paper argues that while the BSA represents an incremental improvement in doctrinal clarity, it does not yet furnish courts with a coherent, technologically adaptive standard for evaluating the authenticity of synthetic or manipulated electronic content. The paper concludes with recommendations for procedural rules under Section 67C of the Information Technology Act, 2000, mandatory forensic standardisation, and a graduated reliability-based test for AI-generated evidence.
KeywordsBharatiya Sakshya Adhiniyam, 2023; digital evidence; electronic records; Section 63 certificate; hash value; admissibility; authentication; Arjun Panditrao Khotkar; deepfake evidence; Indian evidence law.
1. Introduction
The manner in which Indian courts receive and evaluate evidence has been transformed by the near-total migration of communication, commerce, and record-keeping onto digital platforms. Emails, WhatsApp chats, call detail records, CCTV footage, server logs, and cloud-stored data now form the evidentiary backbone of a substantial share of civil and criminal litigation. Until 1 July 2024, this material was governed by Sections 65A and 65B of the Indian Evidence Act, 1872, provisions inserted by the Information Technology Act, 2000, which required a signed certificate as a precondition for admitting “computer output” without producing the original device. The jurisprudence generated by these provisions was notoriously unsettled, oscillating between strict and liberal positions across [1] four major Supreme Court decisions over fifteen years.
The Bharatiya Sakshya Adhiniyam, 2023 [2] (“BSA”) was enacted as part of a wholesale replacement of India’s colonial-era criminal statutes. Insofar as electronic evidence is concerned, the BSA does not discard the certification model built around the erstwhile Section 65B; instead, it relocates the substance of that model into Section 63, adds a new Section 61 that expressly forbids denial of admissibility solely on the ground that a record is electronic, and introduces a materially stricter certification regime under Section 63(4), requiring dual signatures — one from the person in lawful control of the device and a second from an expert — together with disclosure of the record’s hash value. These changes were consciously designed to respond to the rigidity that Arjun Panditrao Khotkar had imposed and to the forensic realities of an era in which digital content can be manipulated with unprecedented ease, including through generative artificial intelligence.
This paper asks three connected questions. First, does Section 61 of the BSA, through its non-obstante-like opening words, restore an alternative, certificate-free route to proving electronic records, or does Section 63 remain, as Arjun Panditrao held of the old Section 65B, a complete and exclusive code? Second, what do the dual-certification and hash-value requirements of Section 63(4) mean in practice for litigants, and how has the Supreme Court’s 2026 decision in Pune Bar Association v. Union of India[3] resolved — and left open — questions surrounding these requirements? Third, is the existing statutory and judicial architecture, built around the paradigm of a static “computer output,” adequate to the authentication challenges posed by deepfakes and other AI-generated content? The paper proceeds by reviewing the relevant literature, analysing the statutory text and case law in detail, presenting findings on the coherence of the current framework, and offering recommendations for legislative and institutional reform.
2. Literature Review
The literature on electronic evidence in India can be organised into three broad strands. The first strand, produced principally in the years following Anvar P.V. and Arjun Panditrao, is doctrinal commentary tracing the vacillation of the Supreme Court on the mandatoriness of the Section 65B(4) certificate. Commentators such as those writing in the India Business Law Journal[4] and in SCC OnLine’s blog observed that Arjun Panditrao resolved the conflict between the three-judge Anvar position and the two-judge Shafhi Mohammad position by holding the certificate mandatory in all cases where the original device is not produced, while creating an equitable escape valve permitting courts to summon the certificate through the court’s process where a party had done everything reasonably possible to obtain it. This body of work is essential background but necessarily predates the BSA and cannot address the 2023 statute’s innovations.
The second strand consists of early commentary on the BSA itself, published between 2024 and 2026 in professional and practitioner outlets. Writers analysing Sections 61 to 63 have converged on the view that Section 63 is the direct successor to Section 65B and “retain[s] its core framework while introducing important refinements, most notably the requirement of dual certification.”[5] A distinct and more contested line of academic commentary, associated principally with Advocate Saji Koduvath, argues that Section 61’s opening words — “Nothing in this Adhiniyam shall apply to deny the admissibility of an electronic or digital record… on the ground that it is an electronic or digital record” — were deliberately enacted to legislatively override the exclusivity holding in Arjun Panditrao, so that a computer output may now be proved by any lawful mode, with Section 63’s certificate route remaining available but no longer mandatory where a witness with direct knowledge is examined.[6] This view remains a minority academic position that has not yet been squarely tested before the Supreme Court, and the present paper engages with it critically in Part 3.
The third strand, concentrated in 2025–2026, addresses the forensic and technological pressure points of the new regime: the practical difficulty of standardising hash-value generation across investigating agencies,[7] the unresolved question of who may sign Part B of the Schedule as an “expert” for the purposes of Section 63(4),[8] and, most urgently, the incapacity of a certification-based admissibility regime to address deepfakes and other synthetic media that can satisfy every procedural requirement of Section 63 while being substantively fabricated.[9] This literature is largely diagnostic rather than prescriptive, and comparative engagement with other jurisdictions — principally the United Kingdom’s Forensic Science Regulator Code of Practice for Digital Evidence and Synthetic Media and the evidentiary provisions of the United States Federal Rules of Evidence 702 and 901 — remains sparse in the Indian context. This paper situates itself at the intersection of the second and third strands, offering a structured doctrinal analysis of Sections 61–63 read with the 2026 Pune Bar Association ruling, and extending the discussion to concrete recommendations.
3. Analysis and Discussion
3.1 The Statutory Architecture: Sections 2, 57, 61, 62 and 63
Section 2(1)(d) of the BSA amends the definition of “document” to expressly include electronic and digital records, thereby eliminating the interpretive gap that had earlier required courts to import electronic records into the concept of “document” by analogy. Section 57 similarly clarifies that electronic records are documentary evidence and may be proved either as primary evidence, where the record itself is produced, or as secondary evidence in the specific manner set out in Section 63. Section 61 then supplies the general non-discrimination principle: an electronic or digital record shall not be denied admissibility solely on the ground that it is electronic, and — critically — such a record shall, subject to Section 63, have the same legal effect, validity and enforceability as any other document.[10] Section 62 restates, in terms mirroring the old Section 65A, that the contents of an electronic record may be proved “in accordance with the provisions of section 63.” Section 63 itself is the operative provision: it deems a computer output — information contained in an electronic record that is printed, stored, recorded, or copied onto optical, magnetic, or semiconductor media, or produced by a computer or communication device — to be a “document,” admissible without further proof of the original, provided the conditions of sub-section (2) are satisfied and a certificate under sub-section (4), in the form prescribed by the Schedule, accompanies it.
The critical textual puzzle lies in the words “subject to section 63” in Section 61. One reading, favoured by most practitioner commentary, treats this qualifying phrase as preserving Section 63’s primacy: Section 61 removes only the in-principle objection that electronic records cannot be documents at all, while the mode of proof for such records remains exclusively governed by Section 63, consistent with Arjun Panditrao’s characterisation of the certificate as a “condition precedent.”[11] A second, minority reading treats the phrase “nothing in this Adhiniyam shall apply to deny admissibility… on the ground that it is an electronic or digital record” as a legislative correction of Arjun Panditrao, restoring the flexibility that Shafhi Mohammad had briefly allowed, on the theory that if Section 61 were read merely as declaratory of Section 63’s primacy it would be otiose, since Section 63 already achieves the same result for records that comply with sub-section (4).[12] On balance, the better view — and the one consistent with the retention of an unqualified certification requirement in Section 63(4) and with the Schedule’s continued prescription of a bifurcated Part A/Part B certificate — is that Section 61 performs a narrower function: it forecloses arguments that electronic records are categorically inadmissible, while Section 63 remains the operative gateway whenever a party seeks to prove a computer output without calling a witness with direct, first-hand knowledge of the record’s creation. Where such a witness is examined and can depose to the making and integrity of the record, general principles of documentary and oral evidence continue to apply, and this has been so both before and after the BSA. The genuine innovation of Section 61 for admissibility purposes is therefore modest, but it is not without consequence: it removes any residual argument, occasionally raised at the margins, that an electronic record is a fundamentally different species of evidence requiring a wholly separate admissibility threshold.
3.2 Section 63(4): The Dual-Certificate and Hash-Value Regime
The most consequential change introduced by the BSA is the restructuring of the Section 65B(4)-style certificate into a two-part instrument under the Schedule to the Act. Part A is completed by the party producing the electronic record or by the person in lawful control of the relevant device or computer system, and must identify the record, describe the manner of its production, and — for the first time in Indian evidence law — disclose the record’s hash value, a fixed-length alphanumeric string generated by a cryptographic algorithm that functions as a unique digital fingerprint of the file’s contents at a given point in time.[13] Part B must additionally be signed by an expert. Section 63(4) does not itself define “expert,” and this gap became the subject of conflicting interpretation almost immediately after the BSA came into force: the Madras High Court took the view that only a person notified as an “Examiner of Electronic Evidence” under Section 79A of the Information Technology Act, 2000 could validly sign Part B, a position that, if correct, would have made compliance practically impossible in most of the country given the limited number of notified examiners.
This precise question reached the Supreme Court in Pune Bar Association v. Union of India, decided on 22 May 2026 by a three-judge bench. The Pune Bar Association had challenged Section 63(4) as unconstitutional on the ground that it imposed an onerous, practically unworkable burden on ordinary litigants by conditioning admissibility on both a hash-value disclosure and an expert signature.[14] The Court rejected the constitutional challenge, holding that the disclosure of a hash value is not an empty technical formality but “an electronic fingerprint” that furnishes courts with an objective, verifiable method of testing whether a file has been tampered with, and that this safeguard is proportionate to the legitimate object of preserving the integrity of electronic evidence in an era of easy manipulation and, increasingly, synthetic media.[15] On the separate question of who may sign as an expert, the Court disagreed with the Madras High Court’s restrictive reading, holding that Sections 39(1) and 39(2) of the BSA, read harmoniously with Section 63(4), permit any person possessing special skill and practical experience in computer science and cyber forensics to sign Part B, and that the category of qualified experts is not confined to persons formally notified under Section 79A of the Information Technology Act, 2000.[16] This holding materially widens the pool of persons competent to certify electronic evidence and addresses, at least in part, the access-to-justice concern that had animated the petition; at the same time, by declining to lay down a closed or codified definition of “expert,” the Court has left considerable discretion to trial courts to assess the sufficiency of a given signatory’s qualifications on a case-by-case basis, which is likely to generate fresh interlocutory litigation over the adequacy of Part B certifications in the years ahead.
3.3 The Arc of Judicial Interpretation: From Navjot Sandhu to Pune Bar Association
The interpretive history that the BSA inherits is essential to understanding why Parliament chose the specific architecture of Sections 61–63. In State (NCT of Delhi) v. Navjot Sandhu[17], arising from the 2001 Parliament attack case, the Supreme Court held that secondary evidence of electronic records, including call records, could be admitted through other modes of proof under Sections 63 and 65 of the Indian Evidence Act, 1872 even without a Section 65B certificate, on the theory that Section 65B was an additional, non-exclusive mode of proof. This liberal position stood for nearly a decade until Anvar P.V. v. P.K. Basheer[18], in which a three-judge bench overruled Navjot Sandhu and held that Sections 65A and 65B constitute a complete and self-contained code for the proof of electronic records, such that oral evidence or a general certificate under Section 63 could not substitute for strict compliance with Section 65B(4). Anvar thus established the principle of mandatory compliance that has, with fluctuations, governed the field since.
The rigour of Anvar produced practical hardship in cases where the certifying party — typically a third-party service provider, bank, or telecom operator — was unwilling or administratively unable to furnish a certificate, prompting a two-judge bench in Shafhi Mohammad v. State of Himachal Pradesh[19] to hold that the Section 65B(4) requirement was procedural rather than substantive, and could be relaxed in the interest of justice where the party adducing the electronic record was not in possession of the device from which it originated. This created a direct doctrinal conflict with Anvar, which the Supreme Court finally resolved in Arjun Panditrao Khotkar v. Kailash Kushanrao Gorantyal[20], an election petition arising from disputed CDs and VCDs used to prove that nomination forms had been filed after the prescribed cut-off time. A three-judge bench, speaking through Nariman J., held Shafhi Mohammad per incuriam, reaffirmed Anvar as correctly decided, and held that a Section 65B(4) certificate is a mandatory condition precedent to the admissibility of a computer output, save that where the certifying authority refuses to cooperate a party may invoke the court’s coercive process, including applications under Sections 165 of the Evidence Act or orders directing production, to secure the certificate or the underlying record. The Court also issued general directions on the preservation of electronic evidence and recommended that rules be framed under Section 67C of the Information Technology Act, 2000 for retention protocols, chain-of-custody documentation, and standardisation across investigating agencies.
The BSA’s Section 63, in retaining a mandatory certificate requirement while adding hash-value disclosure and dual signatures, can be read as Parliament’s codification of the Arjun Panditrao mandatory-compliance position, strengthened rather than relaxed. Pune Bar Association v. Union of India (2026) is, in this sense, the direct doctrinal descendant of Arjun Panditrao: both decisions reject arguments that the certification burden is disproportionate, and both treat rigorous authentication as a necessary safeguard rather than an obstacle to justice. What distinguishes the 2026 decision is its context — a world in which manipulation of digital content, including through generative artificial intelligence, is no longer a marginal risk but a documented and growing feature of litigation, civil and criminal alike, giving the Court’s emphasis on the hash value as an “electronic fingerprint” a materially different weight than similar language would have carried in 2020.
3.4 Authentication Beyond Certification: Deepfakes and AI-Generated Content
A certificate under Section 63(4), however rigorously administered, authenticates the chain of custody and integrity of a computer output — it confirms that the record produced in court is an accurate, unaltered reproduction of what was stored on a given device. It does not, and structurally cannot, authenticate the truth or genuineness of the underlying content where that content was itself synthetically generated. A deepfake video, once created, can be copied, stored, and reproduced through a perfectly regular chain of custody, generating a hash value that verifies only that the copy in court matches the (fabricated) original, not that the original depicts a real event.[21] This is the central structural limitation of a certification-based admissibility regime when confronted with synthetic media, and it has drawn increasing academic and judicial attention through 2025 and 2026.
Indian courts have so far addressed the deepfake problem primarily outside the evidentiary framework, through the law of personality rights and interim injunctive relief rather than through the law of proof. In late 2025, the Delhi High Court and the Bombay High Court granted interim injunctions restraining the circulation of unauthorised deepfakes and AI-generated voice clones of public figures, including well-known film actors, grounding the relief in the personality and publicity rights protected under Article 21 of the Constitution and directing intermediary platforms to take down synthetic impersonations upon notification.[22] These orders address the harm caused by deepfakes as content but do not resolve the distinct evidentiary question of how a court should treat a deepfake, or an ordinary electronic record whose authenticity is challenged on grounds of AI manipulation, once it is tendered as evidence in a trial. The Information Technology Rules, 2026 introduce traceability and labelling obligations for synthetic content, which may in time generate metadata useful for authentication, but these obligations are addressed to platforms and do not by themselves furnish courts with an evidentiary test for content that has evaded labelling or that originates outside a regulated platform.[23]
The absence of a dedicated statutory test creates two distinct risks that the literature has identified: false admission, in which fabricated content satisfies Section 63(4)’s procedural requirements and is admitted despite being substantively unreliable, and false exclusion or the so-called “liar’s dividend,” in which genuine, unfavourable electronic evidence is challenged as a deepfake to erode its credibility even where no forensic basis exists for that challenge.[24] Comparative practice offers only partial guidance: the United Kingdom’s Forensic Science Regulator has issued a Code of Practice for Digital Evidence and Synthetic Media directed at forensic examiners, and courts in the United States have begun to encounter, and struggle with, AI-generated exhibits under the general reliability standards of Federal Rules of Evidence 702 and 901, without a settled consensus on the appropriate threshold.[25] No comparable regulatory or doctrinal standard yet exists in India, and Part 5 of this paper returns to this gap in framing recommendations.
4. Findings
- Section 63 of the BSA is the operative and, in practice, exclusive gateway for admitting a computer output without examining a witness with direct knowledge of the record; Section 61 narrows objections grounded in the electronic character of a record but does not displace Section 63’s certification requirement, notwithstanding a respectable minority academic argument to the contrary.
- The BSA’s dual-certificate, hash-value regime under Section 63(4) is a direct legislative response to the practical difficulties exposed by Arjun Panditrao Khotkar, and its constitutionality — including the hash-value disclosure requirement — has been authoritatively upheld by the Supreme Court in Pune Bar Association v. Union of India (2026).
- The Supreme Court’s clarification that Section 79A-notified examiners are not the only persons competent to sign Part B of the certificate schedule meaningfully widens access to compliant certification, but the absence of a codified definition of “expert” is likely to generate continuing trial-court litigation over the adequacy of individual certifications.
- Hash-value practice remains technically unstandardised: the BSA and its Schedule do not prescribe a mandatory algorithm, verification protocol, or minimum forensic qualification, leaving considerable inconsistency across investigating agencies and forensic laboratories.
- The certification model, whether under the old Section 65B or the new Section 63, authenticates custodial integrity, not substantive truth; it is structurally unable to detect or exclude deepfakes and other AI-generated content that has been properly stored and reproduced, and Indian courts have so far addressed synthetic media chiefly through personality-rights injunctions rather than through the law of evidence.
- No comprehensive Indian statutory or procedural standard yet exists for assessing the reliability of AI-generated or AI-manipulated evidence, in contrast to the more developed, if still unsettled, comparative frameworks emerging under the EU AI Act, the UK Forensic Science Regulator’s Code of Practice, and Federal Rules of Evidence 702 and 901 in the United States.
5. Recommendations
First, the rules contemplated under Section 67C of the Information Technology Act, 2000 and repeatedly urged by the Supreme Court since Arjun Panditrao should be notified without further delay, prescribing mandatory data-retention periods, standardised hash-generation algorithms (at minimum SHA-256 or a demonstrably equivalent successor), and uniform chain-of-custody documentation binding on all investigating agencies and forensic laboratories.
Second, the term “expert” in Section 63(4) and the Schedule should be given content through a notification or practice direction specifying minimum qualifications — whether by reference to Section 79A examiners, recognised certifying bodies, or an accreditation scheme for private cyber-forensic professionals — so as to convert the flexibility recognised in Pune Bar Association into predictable compliance guidance rather than a recurring ground of interlocutory dispute.
Third, the National Informatics Centre, the Central Forensic Science Laboratories, and State Forensic Science Laboratories should be equipped with standardised, validated tools for hash-value verification and for detecting content manipulation, including AI-generated manipulation, with periodic audits to ensure inter-laboratory consistency.
Fourth, given the demonstrated inadequacy of a pure custodial-integrity test for synthetic media, Parliament or the rule-making authority should consider introducing a supplementary, graduated reliability test for electronic records whose authenticity is specifically challenged on grounds of AI generation or manipulation — for instance, requiring the proponent to lead expert evidence on provenance, metadata, and detection-tool analysis once a threshold challenge is raised, broadly analogous to the reliability screening contemplated for expert evidence in comparative systems.
Fifth, judicial and prosecutorial training institutions, including the National Judicial Academy and State Judicial Academies, should incorporate structured training on hash-value verification, forensic certification, and the detection of synthetic media into their curricula, given the pace at which these issues are now reaching trial courts.
Sixth, empirical research tracking trial-court application of Section 63(4) — rates of certificate non-compliance, the profile of persons signing as experts, and the incidence of authenticity challenges based on alleged AI manipulation — should be commissioned, since the existing literature is overwhelmingly doctrinal and lacks an empirical foundation for evaluating whether the 2023 reforms have achieved their intended objectives.
6. Conclusion
The Bharatiya Sakshya Adhiniyam, 2023 consolidates, rather than radically reconceives, the certification-based model of electronic evidence that Indian courts developed under Sections 65A and 65B of the repealed Evidence Act. Sections 61 to 63 clarify the categorical admissibility of electronic records, retain a mandatory certificate as the operative gateway for computer output tendered without a first-hand witness, and respond directly to the practical difficulties exposed across Navjot Sandhu, Anvar, Shafhi Mohammad and Arjun Panditrao by adding hash-value disclosure and dual certification. The Supreme Court’s 2026 decision in Pune Bar Association v. Union of India has now confirmed the constitutional soundness of this stricter regime while easing, though not eliminating, uncertainty over who may certify. These are genuine improvements in doctrinal clarity and forensic rigour. They do not, however, resolve the deeper challenge now facing the law of evidence: a procedural framework built to verify that a copy faithfully reproduces an original cannot, without further reform, verify that the original itself is authentic in an age of generative artificial intelligence. Bridging that gap will require coordinated action across statute, subordinate rule-making, forensic infrastructure, and judicial training — the direction of which this paper has sought to outline.
References
Bharatiya Sakshya Adhiniyam, 2023, No. 47 of 2023 (India).
Indian Evidence Act, 1872 (repealed with effect from 1 July 2024, save as to pending proceedings).
Information Technology Act, 2000, §§ 67C, 79A (India).
Information Technology Rules, 2026 (India).
State (NCT of Delhi) v. Navjot Sandhu, (2005) 11 SCC 600.
Anvar P.V. v. P.K. Basheer, (2014) 10 SCC 473.
Shafhi Mohammad v. State of Himachal Pradesh, (2018) 2 SCC 801.
Tomaso Bruno v. State of Uttar Pradesh, (2015) 7 SCC 178 (overruled in part by Arjun Panditrao Khotkar).
Arjun Panditrao Khotkar v. Kailash Kushanrao Gorantyal, (2020) 7 SCC 1.
Selvi v. State of Karnataka, (2010) 7 SCC 263.
Pune Bar Association v. Union of India, W.P. (Civil) No. 599 of 2026 (S.C., decided 22 May 2026).
Law Commission of India, Report on Digital Evidence and Artificial Intelligence (2024).
“The Supreme Court Brings Consistency to Electronic Evidence,” India Business Law Journal (21 October 2020).
“SC Clarifies Law on Admissibility of Electronic Evidence without Certificate under Section 65B of Evidence Act, 1872,” SCC OnLine Blog (3 December 2020).
“Understanding E-Evidence under Bhartiya Sakshya Adhiniyam 2023: Key Provisions and Implications,” LiveLaw (15 July 2024).
“Electronic Evidence under the BSA, 2023,” iPleaders Blog (2 July 2026).
“Electronic Evidence Changes Will Modernise Banking Practices,” Law.asia (26 February 2026).
KSandK, “Section 63 BSA 2023: Admissibility of Electronic Evidence” (30 May 2026).
Saji Koduvath, “Sections 61, 62, and 63 of the BSA: Admissibility of Electronic Records,” IndianLawLive (2 July 2026).
Saji Koduvath, “‘Nothing in this Adhiniyam Shall Apply to Deny the Admissibility’ – New Provision (S. 61, BSA),” IndianLawLive (10 October 2025).
“Admissibility and Authentication of Electronic Evidence under the Bharatiya Sakshya Adhiniyam, 2023: Legal Framework and Emerging Challenges,” Mondaq (24 June 2026).
JSA, “JSA Prism | Dispute Resolution” (June 2026).
ThanggoulenHaokip, “Who Can Certify Digital Evidence? India’s Supreme Court Leaves the Answer Open,” Legal Service India (20 June 2026).
“Law Web: Supreme Court Rejects Challenge to S.63(4) BSA Mandating Hash Value Disclosure for Electronic Evidence” (22 May 2026).
Bhatt & Joshi Associates, “Importance of Hash Value in the Context of Digital Evidence Collection” (5 May 2026).
“Digital Evidence in Criminal Trials: Section 63 of the Bharatiya Sakshya Adhiniyam, Hash Values & the New Rules of Admissibility,” LawBeat (2026).
“Deepfakes as Evidence: Authenticity, Chain of Custody and Section 65B in the Age of IT Rules 2026,” Mondaq (24 March 2026).
“Deepfake Evidence in Indian Courts: Legal Challenges, Admissibility, Authentication & Burden of Proof,” Legal Service India (4 July 2026).
Vansh Srivastava & Arvind Kumar Singh, “Legal Status of AI-Generated Evidence in Indian Courts,” LIJDLR, Vol. 4 (2026).
“Can Courts Trust Deepfakes? Reimagining Electronic Evidence in the AI Era,” Jus Corpus (2026).
“Deepfake Evidence and the Indian Criminal Justice System,” International Journal for Multidisciplinary Research (2025).
“Deepfakes and False Digital Evidence in Indian Courts,” International Journal of Legal Studies, Vol. 12, Issue 1 (2026).
“AI-Generated Evidence in Indian Courts: Admissibility, Reliability, and the Chain-of-Custody Challenge,” IJIRL (September 2025).
UK Forensic Science Regulator, Code of Practice for Digital Evidence and Synthetic Media (Home Office, 2023).
Regulation (EU) 2024/1689 (EU Artificial Intelligence Act), art. 50.
United States Federal Rules of Evidence, rr. 702, 901.
Jared Perlo, “AI-Generated Evidence Is Showing Up in Court. Judges Say They’re Not Ready,” NBC News (18 November 2025).
[1]State (NCT of Delhi) v. Navjot Sandhu, (2005) 11 SCC 600; Anvar P.V. v. P.K. Basheer, (2014) 10 SCC 473; Shafhi Mohammad v. State of Himachal Pradesh, (2018) 2 SCC 801; Arjun Panditrao Khotkar v. Kailash Kushanrao Gorantyal, (2020) 7 SCC 1.
[2]The Bharatiya Sakshya Adhiniyam, 2023, No. 47 of 2023 (India), received presidential assent on 25 December 2023 and, together with the Bharatiya Nyaya Sanhita, 2023 and the Bharatiya Nagarik Suraksha Sanhita, 2023, came into force on 1 July 2024 vide S.O. 2673(E).
[3]Pune Bar Association v. Union of India, W.P. (Civil) No. 599 of 2026 (S.C., decided 22 May 2026).
[4]“The Supreme Court Brings Consistency to Electronic Evidence,” India Business Law Journal (21 October 2020).
[5]Khaitan & Co./KSandK, “Section 63 BSA 2023: Admissibility of Electronic Evidence” (30 May 2026).
[6]Saji Koduvath, “‘Nothing in this Adhiniyam shall apply to deny the admissibility’ – New Provision (S. 61, BSA) to Ensure that S. 63, BSA (S. 65B, Evidence Act) is an Enabling Provision,” Indian Law Live (10 October 2025).
[7]Bhatt & Joshi Associates, “Importance of Hash Value in the Context of Digital Evidence Collection” (5 May 2026).
[8]ThanggoulenHaokip, “Who Can Certify Digital Evidence? India’s Supreme Court Leaves the Answer Open,” Legal Service India (20 June 2026).
[9]Vansh Srivastava & Arvind Kumar Singh, “Legal Status of AI-Generated Evidence in Indian Courts,” LIJDLR (2026); see also “Can Courts Trust Deepfakes? Reimagining Electronic Evidence in the AI Era,” Jus Corpus (2026).
[10]Bharatiya Sakshya Adhiniyam, 2023, § 61.
[11]Arjun Panditrao Khotkar v. Kailash Kushanrao Gorantyal, (2020) 7 SCC 1, para 24: “The general directions issued in para 62 of Anvar P.V. (supra) shall hereafter be followed by courts that deal with electronic evidence, to obviate any confusion.” See also para 21, holding the Section 65B(4) certificate a “condition precedent” to admissibility.
[12]Saji Koduvath, supra note 8.
[13]See generally Bhatt & Joshi Associates, supra note 9, describing hash values as enabling a court to “compare[] the hash value of received data to the hash value of the original data to check if it has been altered.”
[14]Pune Bar Association v. Union of India, W.P. (Civil) No. 599 of 2026, para 1 (S.C., 22 May 2026).
[15]JSA Prism, “Dispute Resolution” (June 2026), summarising the Court’s characterisation of the hash value as a substantive safeguard rather than a mere formality.
[16]JSA Prism, supra note 12.
[17]State (NCT of Delhi) v. Navjot Sandhu, (2005) 11 SCC 600 (the Parliament attack case).
[18]Anvar P.V. v. P.K. Basheer, (2014) 10 SCC 473.
[19]Shafhi Mohammad v. State of Himachal Pradesh, (2018) 2 SCC 801.
[20]Arjun Panditrao Khotkar v. Kailash Kushanrao Gorantyal, (2020) 7 SCC 1.
[21]See “Deepfakes as Evidence: Authenticity, Chain of Custody and Section 65B in the Age of IT Rules 2026,” Mondaq (24 March 2026), noting that deepfakes are treated as electronic records subject to the ordinary Section 63 (formerly 65B) framework, which was not designed to test authenticity of content.
[22]“Deepfake Evidence in Indian Courts: Legal Challenges, Admissibility, Authentication & Burden of Proof,” Legal Service India (4 July 2026), citing interim orders obtained in December 2025 by several film actors against circulation of deepfake videos and voice clones.
[23]Mondaq, supra note 26.
[24]Vansh Srivastava & Arvind Kumar Singh, supra note 10; Jus Corpus, supra note 11, discussing the erosion of the evidentiary intuition that “seeing is believing.”
[25]UK Forensic Science Regulator, Code of Practice for Digital Evidence and Synthetic Media (2023); see also Jared Perlo, “AI-Generated Evidence Is Showing Up in Court. Judges Say They’re Not Ready,” NBC News (18 November 2025).