Abstract
The way courts collect, analyse, and present evidence is being altered by artificial intelligence (AI). Traditionally, courts relied on human witnesses, document-based evidence, and expert testimony to build their case. Recently, however, the introduction of machine learning (ML), facial recognition (FR), predictive analytics (PA), digital forensics (DF) and automated surveillance (AS) has resulted in AI-based evidence being included in the court system. The research will utilise both doctrinal and analytical methods to determine how AI-based evidence can be admitted, valued as evidence, and given a status under the Bharatiya Sakshya Adhiniyam, 2023 (the Evidence Act). The study also examines how AI is used during criminal investigations; how to detect cybercrime; how to perform forensic examinations; and how to manage evidence. Another reason this study is of great significance is that it contains information on how AI will improve the accuracy, efficiency and reliability of processing complex digital data while minimising the potential for human errors to be made. There are still many barriers to overcome with respect to algorithmic bias, transparency, data privacy, accountability and the integrity of AI-generated evidence. The findings of the study indicate that AI is to be used in conjunction with human judgement in the justice system. The study provides valuable insight into the intersection between law and technology for those studying or otherwise interested in evidence and justice as it relates to the digital world. The study reaches one additional conclusion: that there is a pressing need for a comprehensive legal and ethical framework that will facilitate the responsible integration of AI into the justice system.
KEYWORDS: Artificial Intelligence (AI), AI-Generated Evidence, Bharatiya Sakshya Adhiniyam, 2023, Digital Forensics, Admissibility of Evidence, Criminal Investigation, Algorithmic Bias, Administration of Justice.
INTRODUCTION
Traditionally, the judicial system relied heavily on humans to provide information about events, and as a result, the provision of that information in the form of witness testimony was considered the most important way to establish a case legally. However, in most cases, witnesses were biased against one another for various reasons. As a result, they may misremember events because of threatening or intimidating circumstances or their flawed ability to recall what truly occurred.
As courts started to realise the limitations of relying upon human testimony to establish factual determinations and dispute resolution, they became more willing to accept scientific and digital evidence as a way to establish those same factual determinations and provide a resolution to those disputes. Technology, such as computers, cell phones and emails, has made documenting events much easier, which contributes to the courts’ acceptance of evidence because it provides the courts with more accurate accounts of events. Therefore, digital evidence plays a critical role in the judicial system’s ability to meet its burden of proof to establish a case legally. Due to the increased volume of digital evidence, the courts have begun to develop laws and rules of evidence related to these forms of evidence.
Currently, courts are beginning to accept AI as a form of evidence for establishing fact and providing resolution to a dispute. However, in contrast to traditional digital evidence, AI can access and analyse large volumes of data, identify patterns, assist with facial recognition technology, support digital forensics, and produce more efficient investigations than human technology alone can provide. In addition, courts will be able to use AI to make judicial determinations faster, improve the accuracy of the determinations, and ultimately create a more equitable, fair and just judicial process, which would constitute a major improvement over human testimony or traditional digital evidence.
This study examines the nature, scope, and evidentiary value of AI-generated evidence, focusing on its distinction from traditional forms of evidence such as oral testimony, documentary records, and expert opinions. It critically analyses the legal challenges surrounding the admissibility, reliability, authenticity, transparency, and accountability of AI-generated evidence in judicial proceedings. The research further evaluates whether the existing Indian legal framework, particularly the Bharatiya Sakshya Adhiniyam, 2023, is adequately equipped to address the growing integration of AI within the justice system and identifies reforms necessary for its effective regulation. Adopting a doctrinal research methodology, the study relies on statutes, judicial precedents, scholarly literature, policy reports, and comparative legal analysis. It examines landmark decisions including Anvar P.V. v. P.K. Basheer[1], Arjun Panditrao Khotkar v. Kailash Kushanrao Gorantyal[2], and Justice K.S. Puttaswamy v. Union of India[3], while drawing insights from international legal approaches to AI evidence.
The Bharatiya Sakshya Adhiniyam, 2023[4] and the Information Technology Act, 2000[5]recognise electronic records and digital proof, but do not meet the demands created by AI-created proof in the law. Both rules were designed primarily to provide a method of regulating electronic records, confirming their admission into court as well as any lawful establishment of the proof’s reliability, value and authenticity. However, there are no guidelines contained within these laws to specify different treatments for the output of AI systems such as machine learning algorithms, predictive analytics, autonomous decision-making systems and deepfakes; therefore,several basic principles regarding how to assess algorithmic transparency, explainability, accountability, reliability and threats of bias associated with these tools have no regulatory framework. The increasing presence of AI in such areas as facial recognition, voice analysis, forensic investigations and assessment of evidence makes the lack of a specific statutory framework surrounding the admissibility, weight of evidence, liability for errors or infringement of fundamental rights difficult to navigate legally. The need to create a regulatory scheme to cover the use of AI-generated proof requires comprehensive reform to ensure a fair, transparent, due process and properly judicially supervised administration of justice.
[1]Anvar P.V. v. P.K. Basheer, (2014) 10 S.C.C. 473 (India).
[2]Arjun Panditrao Khotkar v. Kailash Kushanrao Gorantyal, (2020) 7 SCC 1 (India).
[3]Justice K.S. Puttaswamy (Retd.) v. Union of India, (2017) 10 S.C.C. 1 (India).
[4]Bharatiya Sakshya Adhiniyam, No. 47 of 2023, §§ 61–63 (India).
[5]Information Technology Act, No. 21 of 2000, § 4, India Code (2000) (recognizing the legal validity of electronic records).