Deepfake Evidence India: Admissibility, Authenticity, and the Future of Digital Proof in Indian Courts

Say you are a judge. A lawyer submits a video clip as evidence. In it, a man clearly admits to a crime. The clip looks real. The voice sounds right. The face matches. But here is the problem: the clip was made by an AI. The man never said those words. He never sat in that room. The video is a deepfake, and it is sitting in your courtroom as "proof."

This is not a hypothetical ripped from a Black Mirror episode. It is a very real and growing challenge facing Indian courts today. As deepfake technology becomes cheaper and more accessible, the line between authentic digital evidence and sophisticated forgery is getting harder to see with the naked eye. For lawyers, judges, and anyone involved in litigation, understanding how Indian law currently handles deepfake evidence in India is no longer optional. It is essential.

What Exactly Is a Deepfake?

Before diving into law, let us get the technology right. A deepfake is a piece of media, typically a video, audio recording, or image, generated or manipulated using artificial intelligence. The term comes from combining "deep learning" (a type of AI technique) with "fake." The AI is trained on real footage of a person and then used to convincingly replace or alter their face, voice, or both.

The results can be startling. Early deepfakes were easy to spot: the blinking was off, the skin tone did not match, or the audio felt slightly out of sync. Today, state-of-the-art deepfakes can fool both ordinary people and, alarmingly, forensic investigators who are not equipped with the right tools.

In India, deepfakes have already surfaced in public discourse. From viral manipulated videos of politicians to non-consensual intimate imagery involving celebrities, the harm is documented. What is less discussed is how this technology is quietly entering legal disputes, from criminal cases and matrimonial litigation to corporate fraud and property disputes.

The Legal Framework: Where Does India Stand?

India does not yet have a dedicated law on deepfakes. That gap matters enormously when it comes to admissibility of AI-generated evidence in India. So, which laws currently apply?

The Indian Evidence Act, 1872 (and Its Digital Update)

The primary law governing evidence in India is the Indian Evidence Act, 1872. Sections 65A and 65B were added through the Information Technology Act, 2000, to deal with electronic records specifically. Section 65B is the cornerstone of digital evidence admissibility in India. It requires that anyone wishing to produce electronic evidence must provide a certificate confirming that:

• The electronic record was produced by a computer during its regular operation.

• The computer was functioning properly at the time.

• The information in the record was regularly fed into the computer.

The Supreme Court, in the landmark case of Arjun Panditrao Khotkar v. Kailash Kushanrao Gorantyal (2020), made the Section 65B certificate mandatory for the admissibility of electronic records. No certificate, no admissibility. Period.

Now here is the problem with deepfakes: a Section 65B certificate confirms that a file came from a specific device or system. It does not confirm that the content of that file is authentic or unmanipulated. A deepfake video can very easily satisfy the technical requirements of Section 65B while still being completely fabricated. The certificate proves origin, not truth.

The Bharatiya Sakshya Adhiniyam, 2023

India's new evidence law, the Bharatiya Sakshya Adhiniyam (BSA) 2023, which came into force in July 2024 and replaced the Indian Evidence Act, expands the definition of electronic records to include data generated by AI systems. This is a meaningful step forward. However, the BSA still lacks specific provisions dealing with deepfake evidence in India or mechanisms to verify AI-generated content. The authentication framework remains largely the same, certificate-based, without mandatory forensic verification of content integrity.

The Information Technology Act, 2000

Section 66D of the IT Act penalises impersonation using communication devices. Section 66E protects privacy. Section 67 and 67A deal with obscene and sexually explicit content. Some of these provisions can be stretched to cover deepfake-related offences. But stretched is the operative word. They were written for a pre-deepfake era and do not squarely address the creation or use of AI-generated media as false evidence.

Admissibility of AI-Generated Evidence in India: The Core Problem

Here is where things get legally interesting. When a party submits a video or audio clip as evidence, the court asks two fundamental questions: Is this evidence relevant? Is this evidence authentic?

Relevance is usually the easier question. Authenticity, in the age of deepfakes, is where the system struggles.

Under existing Indian law, the party submitting evidence has the burden of proving it is genuine if the other side raises a reasonable challenge. But how is a lawyer or judge supposed to evaluate digital evidence authenticity when the forgery is AI-generated and indistinguishable from real footage without specialised tools?

Indian courts currently rely on expert witnesses for technical questions. Under Section 45 of the Indian Evidence Act (and its BSA equivalent), courts can call upon "experts" in science, art, or foreign law. Forensic digital experts can be appointed to examine evidence. However, the capacity of Indian forensic labs to detect sophisticated AI-generated evidence in India is still developing. Many district courts and even some High Courts do not have standardised protocols for deepfake detection. The field is moving faster than the institutions that need to keep up with it.

This creates a dangerous asymmetry. A well-funded party can hire private forensic experts with cutting-edge detection tools. A party with fewer resources may not have that option. And judges, without technical training, are often left to weigh competing expert opinions without clear guidance.

Real-World Scenarios Where Deepfake Evidence in India Could Surface

Let us look at concrete situations where this becomes legally critical.

Matrimonial Disputes: A spouse produces a video that appears to show their partner in a compromising or adulterous situation. If the video is a deepfake, it could wrongfully tilt a custody battle or divorce settlement.

Criminal Trials: A deepfake alibi video could show an accused person in a different location at the time of the crime. Conversely, a fabricated confession video could be used to frame an innocent person.

Corporate Fraud: A doctored audio clip of a CEO issuing instructions could be used in disputes over insider trading, breach of contract, or financial fraud.

Political and Electoral Interference: A deepfake of a political candidate saying something inflammatory, submitted as evidence in an election petition, could have far-reaching democratic consequences.

In each of these cases, the question of digital evidence authenticity is not abstract. It has direct consequences for justice.

How Can Courts Verify Digital Evidence Authenticity?

So, what tools and methods exist to separate real digital evidence from AI-generated fakes?

Forensic Analysis

Deepfake detection software analyses micro-inconsistencies in video and audio that the human eye misses. These include unnatural blinking patterns, subtle skin texture anomalies, inconsistencies in lighting and shadow, and audio-visual mismatches. Tools developed by companies and academic labs can flag content with a probability score of being AI-generated.

However, these tools are not foolproof. As detection improves, so does generation. It is an ongoing arms race.

Blockchain-Based Provenance

Some digital platforms are beginning to embed cryptographic metadata into original content at the point of capture. This creates an unalterable chain of custody for the file. If evidence submitted in court lacks this provenance data, or if the metadata has been tampered with, it is a red flag.

Hash Value Verification

Every digital file has a unique cryptographic hash, essentially a digital fingerprint. If a file has been altered even slightly, its hash changes. Comparing the hash of a submitted file against a verified original (where one exists) can confirm or deny tampering.

Cross-Referencing with Other Evidence

Deepfakes rarely exist in a vacuum. Corroborating or contradicting other evidence, such as location data, communication records, independent witness accounts, and bank transactions, can contextualise and test the credibility of disputed digital content.

What Indian Courts and Lawmakers Need to Do

The current legal framework for handling AI-generated evidence in India has real gaps. Addressing them requires action on multiple fronts.

Mandatory Forensic Protocols: Courts, especially in high-stakes criminal and civil cases, should require forensic verification of digital evidence whenever authenticity is challenged. The Supreme Court or Law Commission could issue guidelines standardising this process.

Accreditation of Deepfake Detection Labs: India needs a credentialled network of forensic labs capable of examining AI-generated content. The Central Forensic Science Laboratory (CFSL) and state counterparts need investment and training in deepfake detection tools.

Updating the Bharatiya Sakshya Adhiniyam: The BSA is new, but it could already benefit from an amendment or specific rules addressing the admissibility and verification of AI-generated evidence in India. Even a detailed rule-making provision, allowing courts or ministries to prescribe technical standards, would help.

Judicial Training: Judges need exposure to how deepfakes work and what current detection limitations look like. This does not mean turning judges into technologists. It means giving them a baseline to ask the right questions and evaluate expert testimony more critically.

Shifting the Burden of Proof Strategically: Where there is credible evidence that a digital submission may be AI-generated, courts should consider placing the burden on the party submitting it to affirmatively prove authenticity through forensic means, rather than requiring the challenging party to prove it is fake.

The Bigger Picture: Trust in the Evidentiary System

There is something more at stake here than legal technicality. Courts run on trust. When parties walk into a courtroom, there is an expectation that evidence, however disputed, represents reality. Deepfakes threaten to corrode that expectation at its foundation.

If litigants and the public begin to doubt whether any digital evidence can be trusted, the legitimacy of verdicts based on such evidence becomes questionable. This erosion of trust is perhaps the most dangerous long-term consequence of unchecked deepfake evidence in India's legal system.

AI-powered legal research tools can help lawyers quickly surface the latest judgments, statutory provisions, and academic commentary on digital evidence standards. They can support lawyers in framing the right objections, identifying relevant precedents, and building arguments that challenge or authenticate digital evidence effectively.

The law will always lag behind technology. That is the nature of the beast. But the gap between legal frameworks and technological reality does not have to be as wide as it currently is. With smarter tools, better-trained professionals, and more agile lawmaking, Indian courts can meet the deepfake challenge without sacrificing the principles that make evidence law work.

Conclusion

Deepfakes are not just a technology problem or a social media nuisance. They are a direct challenge to the evidentiary foundations of justice. Indian law, as it stands, is not fully equipped to handle AI-generated evidence or verify digital evidence authenticity in a standardised, reliable way. The Bharatiya Sakshya Adhiniyam is a starting point, but it is not a finish line.

For lawyers, the immediate takeaway is clear: when digital evidence is involved, ask hard questions about provenance, raise authenticity objections early, and demand forensic verification. For courts, the challenge is to develop the institutional capacity to handle deepfake evidence in India with the rigour it demands. And for lawmakers, the window to act proactively, before deepfake-related miscarriages of justice become commonplace, is still open, but it will not stay open forever.

FAQs

Q1. Can Indian courts admit deepfake videos or AI-generated content as evidence?

Yes, Indian courts can potentially admit AI-generated or manipulated digital content if it satisfies legal requirements for electronic evidence. However, admissibility alone does not establish authenticity. Even if a digital record complies with procedural requirements like certification, courts may still examine whether the content itself is genuine or has been altered through AI tools such as deepfakes.

Q2. Does a Section 65B certificate prove that digital evidence is authentic?

No. A Section 65B certificate primarily establishes the source and manner in which an electronic record was produced. It helps satisfy admissibility requirements for electronic evidence, but it does not guarantee that the content is genuine or unmanipulated. A deepfake file may satisfy technical certification requirements while still containing fabricated content.

Q3. How can courts detect or verify deepfake evidence in India?

Courts may rely on digital forensic examination and expert testimony to evaluate authenticity. Methods such as deepfake detection software, metadata examination, hash-value verification, blockchain provenance checks, and corroboration with surrounding evidence can help determine whether digital evidence has been manipulated.

Q4. Why does deepfake evidence pose a serious challenge to the Indian justice system?

Deepfakes create uncertainty around whether digital evidence reflects reality. In criminal cases, family disputes, corporate litigation, or election matters, fabricated audio or video can influence judicial outcomes if authenticity is not rigorously examined. The broader concern is that widespread uncertainty around digital proof could weaken trust in the evidentiary process itself.