ARTIFICIAL INTELLIGENCE AND EVIDENCE LAW IN PAKISTAN: ADMISSIBILITY, CHALLENGES, AND JUDICIAL APPROACHES

Abstract

Research Background: The rapid proliferation of Artificial Intelligence (AI) and automated algorithmic outputs in criminal, civil, and corporate domains has created a profound tension within traditional legal systems worldwide. In Pakistan, the procedural mechanisms for evaluating evidence remain anchored in conventional statutory frameworks.

Research Problem: The core legal architecture—governed by the Qanun-e-Shahadat Order, 1984 (QSO), the Electronic Transactions Ordinance, 2002 (ETO), and the Prevention of Electronic Crimes Act, 2016 (PECA)—was designed for human-generated electronic records, leaving severe statutory gaps regarding autonomous algorithmic outputs, generative deepfakes, and machine learning inferences.

Objectives: This study evaluates the legal admissibility of AI-generated evidence under Pakistani law, identifies procedural bottlenecks, critically assesses judicial precedents, and offers statutory and technical reform frameworks.

Methodology: The paper utilizes doctrinal legal research, qualitative statutory interpretation, and a comparative legal analysis involving the United States, the United Kingdom, and the European Union.

Key Findings: Pakistani courts currently lack standardized guidelines for evaluating the authenticity, explainability ("black box"), algorithmic bias, and chain of custody for autonomous AI artifacts, creating substantial risks under Constitutional guarantees of fair trial.

Recommendations: Legislative amendments to Article 164 of the QSO, technical forensic validation protocols, judicial training modules at the Federal Judicial Academy, and an AI certification framework are urgently required.

Original Contribution: This research establishes the first comprehensive statutory and jurisprudential framework tailored specifically for integrating AI evidence into Pakistan's evidentiary regime while upholding Article 10-A constitutional rights.