The Illusion of Truth: Navigating Deepfakes and AI-Generated Evidence in Pakistani Courts

The landscape of litigation in Pakistan is undergoing a silent but seismic shift. Imagine a family dispute where an audio recording surfaces, features a distinct voice, and seemingly settles a property claim. Or picture a high-stakes corporate fraud case where a leaked CCTV clip explicitly implicates a Chief Executive. In the past, such electronic evidence was considered a "smoking gun." Today, in the era of generative Artificial Intelligence (AI), it can easily be a complete illusion.

Deepfakes, highly realistic, AI-manipulated, or completely fabricated video, audio, and imagery, have transitioned from a high-tech novelty to a tangible threat within the Pakistani legal framework. For litigators, corporate executives, and ordinary citizens seeking justice, the question is no longer just whether evidence is relevant, but whether it is real.

The Double-Edged Sword of Digital Evidence

The human brain is naturally wired to believe what it sees and hears. When an audio tape or video clip is played in a courtroom, it carries an immediate emotional and persuasive weight. However, as generative AI tools become universally accessible, creating a hyper-realistic deepfake requires little more than a smartphone and a few minutes of source material.

This reality introduces two profound crises into contemporary Pakistani litigation:

The Fabricated Truth: Fraudulent actors can fabricate sophisticated digital evidence, such as fake audio recordings or manipulated video footage to falsely implicate adversaries in civil, criminal, or family matters.

The Liar’s Dividend: Conversely, genuinely guilty parties can now exploit the mere existence of deepfakes to cast doubt on authentic, damning digital evidence, simply claiming, "That video is an AI-generated fake."

The Statutory Landscape: Where Pakistani Law Stands

Pakistan’s legal system possesses an evolving statutory framework to address electronic data, yet it faces unique friction when confronting advanced AI manipulation.

1. The Admissibility Threshold: Qanoon-e-Shahadat Order, 1984 (QSO)

The bedrock of evidence law in Pakistan, the Qanoon-e-Shahadat Order, 1984, was modernized via the Electronic Transactions Ordinance (ETO), 2002, to accommodate the digital age.

• Article 164 of the QSO empowers courts to receive evidence generated through modern devices or techniques. This is the primary gateway through which audio recordings, videos, and WhatsApp logs enter the court record.
• However, admissibility does not equal weight. Under Articles 73 and 78A, the court must be satisfied regarding the integrity, chain of custody, and authenticity of the electronic record. The burden of proving that a digital document has not been altered rests squarely on the party producing it.

2. The Penal Framework: Prevention of Electronic Crimes Act, 2016 (PECA)

Where the QSO governs how evidence is looked at, PECA criminalizes the malicious creation of deepfakes.

• Section 20 (Offences Against Dignity of Natural Person): This section criminalizes the transmission of false information that harms the reputation or privacy of a person, a provision frequently invoked in cyber-defamation cases involving morphed images or deepfake videos.
• Section 13 & 14 (Electronic Forgery and Fraud): Creating a deepfake to illegally alter property rights, alter corporate shares, or secure financial gain constitutes sophisticated electronic forgery.

The Judicial Benchmark

In the landmark judgment ISHTIAQ AHMED MIRZA vs  FEDERATION OF PAKISTAN (P L D 2019 Supreme Court 675), the Supreme Court of Pakistan laid down strict prerequisites for the reliance on audio and video evidence. The court held that digital evidence is inadmissible unless its authenticity is verified, the source device is traced, and the possibility of tampering or editing is conclusively ruled out. In the age of AI, this precedent has become a crucial defense mechanism.

The Honorable Supreme Court of Pakistan has categorically defined the following requirements for admissibility of an audio tape or video in evidence before a court of law and the mode and manner of proving the same before the court:

(i) No audio tape or video could be relied upon by a court until the same was proved to be genuine and not tampered with or doctored.

(ii) A forensic report prepared by an analyst of the Provincial Forensic Science Agency in respect of an audio tape or video was per se admissible in evidence in 2007.

(iii) Under Article 164 of the Qanun-e-Shahadat Order, 1984, it is at the discretion of a court to allow any evidence becoming available through an audiotape or video to be produced.

(iv) Even where a court allowed an audio tape or video to be produced in evidence, such audio tape or video had to be proved in accordance with the law of evidence.

(v) Accuracy of the recording must be proved and satisfactory evidence, director circumstantial, had to be produced so as to rule out any possibility of tampering with the record.

(vi) An audio tape or video sought to be produced in evidence must be the actual record of the conversation as and when it was made or of the event as and when it took place.

(vii) The person recording the conversation or event had to be present.

(viii) The person recording the conversation or event must produce the audiotape or video himself.

(ix) The audio tape or video must be played in the court.

(x) An audio tape or video produced before a court as evidence ought to be clearly audible or viewable.

(xi) The person recording the conversation or event must identify the voice of the person speaking or the person seen, or the voice or person seen may be identified by any other person who recognized such voice or person.

(xii) Any other person present at the time of making of the conversation or taking place of the event may also testify in support of the conversation heard in the audio tape or the event shown in the video.

(xiii) The voices recorded or the persons shown must be properly identified.

(xiv) The evidence sought to be produced through an audio tape or video had to be relevant to the controversy and otherwise admissible.

(xv) Safe custody of the audio tape or video after its preparation till production before the court must be proved.

(xvi) The transcript of the audio tape or video must have been prepared under independent supervision and control.

(xvii) The person recording an audio tape or video may be a person whose part of routine duties was recording an audio tape or video, and he should not be a person who has recorded the audio tape or video for the purpose of laying a trap to procure evidence.

(xviii) The source of an audio tape or video becoming available had to be disclosed.

(xix) The date of acquiring the audio tape or video by the person producing it before the court ought to be disclosed by such person.

(xx) An audio tape or video produced at a late stage of a judicial proceeding may be looked at with suspicion.

(xxi) A formal application had to be filed before the court by the person desiring an audio tape or video to be brought on the record of the case as evidence.

Combatting AI Illusions: The Technical and Legal Defense

Proving that a piece of digital evidence is a deepfake requires a deliberate synthesis of advanced forensics and strategic lawyering. Traditional forensic techniques, which look for basic digital editing or simple file metadata anomalies, often fail against sophisticated neural-network manipulations.

To successfully challenge or defend digital evidence in a Pakistani court, three elements must be rigorously pursued:

1. Demanding an Expert Forensic Audit

Under Section 510 of the Criminal Procedure Code (CrPC) and relevant provisions of PECA, courts can seek opinions from designated forensic experts, such as the Federal Investigation Agency's (FIA) Cyber Crime Wing or specialized digital forensics laboratories. A deepfake can be unmasked by analyzing biological inconsistencies (e.g., unnatural blinking patterns, irregular blood flow reflections on skin) and acoustic physics (e.g., background noise that does not match the room environment).

2. Establishing the Chain of Custody

A litigant must rigorously cross-examine the opposing party on the origin of the digital media. Where was the original device? Who recorded it? How was the file transferred? Any gap in the chain of custody creates reasonable doubt, rendering the evidence highly unreliable under Article 164 of the QSO.

3. Deploying Modern Solutions: Neuro-Symbolic AI

The global legal landscape is shifting toward advanced technological verification. Solutions like Neuro-Symbolic AI combine deep learning with rule-based logic to evaluate digital media. Instead of just identifying pixel anomalies, it assesses whether the context, physical movements, and audio frequencies align naturally, providing clear, auditable reports that judges can rely upon to make informed decisions.

Securing Your Legal Position in a Digital World

As AI tools grow more sophisticated, the line between reality and fabrication will continue to blur. For corporations shielding their assets and individuals protecting their reputations in Pakistan, proactive legal engineering is vital.

Conclusion

When digital evidence threatens your case or business integrity, navigating the complex intersection of technology and law requires specialized expertise. Securing proper forensic intervention early and building a robust evidentiary strategy under the QSO and PECA are your best defenses against digital manipulation.

Disclaimer: This article is intended for informational and educational purposes only and does not constitute formal legal advice. For specialized counsel regarding cyber crimes, digital evidence verification, or corporate litigation within Pakistan, please contact our legal team directly.