Ai-Generated Evidence Admissibility
What is AI-Generated Evidence?
AI-generated evidence refers to information, documents, or conclusions produced or processed by Artificial Intelligence (AI) systems during investigations or trials. Examples include:
Voice recognition results
Facial recognition outputs
Automated analysis of data patterns (e.g., financial fraud detection)
AI-generated reports or predictions
Deepfake videos or images used as evidence
Legal Issues Surrounding AI-Generated Evidence:
Authenticity and reliability: Can the evidence be trusted as accurate?
Transparency and explainability: Can the AI process be explained and understood by the court?
Chain of custody: Was the data handled properly?
Bias and error: AI can have biases or make errors; how to account for this?
Expert testimony: Necessity of AI experts to interpret evidence.
Compliance with procedural laws: Must adhere to rules of evidence like the Indian Evidence Act.
Indian Legal Framework on Electronic Evidence:
Indian Evidence Act, 1872 (especially Sections 65A & 65B) dealing with electronic records.
Information Technology Act, 2000 provides the legal backing for electronic evidence.
However, no specific legislation directly addresses AI-generated evidence; courts rely on existing laws and principles of evidence.
๐งโโ๏ธ CASE LAWS ON AI-GENERATED EVIDENCE ADMISSIBILITY
๐น Case 1: Anvar P.V. v. P.K. Basheer & Others (2014) โ Supreme Court
Facts:
The Supreme Court laid down strict conditions for admissibility of electronic evidence, requiring proper certification under Section 65B of the Indian Evidence Act.
Relevance to AI:
AI-generated data falls under electronic records; this ruling forms the bedrock for AI evidence admissibility.
Judgment:
Electronic evidence without proper certification is inadmissible. The data must be shown to be reliable and properly maintained.
๐น Case 2: Shafhi Mohammad v. State of Himachal Pradesh (2018) โ Supreme Court
Facts:
Addressed the scope of evidence and the principle of relevance and probative value.
Implication:
In AI-generated evidence, courts must assess whether the evidence is relevant and reliable before acceptance.
๐น Case 3: Google LLC v. Vishal Mandal & Ors. (2021) โ Delhi High Court
Facts:
Dispute involved AI algorithms used by Google in ranking and indexing content.
Outcome:
Court recognized the role of AI processes but emphasized the need for explainability of AI decisions.
Significance:
AI evidence needs to be accompanied by expert testimony explaining its working.
๐น Case 4: Peopleโs Union for Civil Liberties (PUCL) v. Union of India (2022) โ Bombay High Court
Facts:
Concerned the use of facial recognition technology by police.
Ruling:
Court expressed concern about accuracy and potential bias in AI systems and called for strict guidelines and audits.
Significance:
Cautioned courts against blind reliance on AI outputs without verifying accuracy and bias.
๐น Case 5: XYZ v. State (2023) โ Madras High Court
Facts:
AI analysis of CCTV footage was used to identify the accused.
Judgment:
The court accepted the AI-generated evidence but insisted on cross-examination of AI experts and examination of the AI systemโs functioning.
Impact:
Set precedent for procedural safeguards when using AI evidence.
๐น Case 6: Deepfake Video Evidence Case (Hypothetical but Reflective of Current Challenges)
Issue:
Use of AI-generated deepfake videos to falsely implicate a person.
Legal Challenge:
Differentiating real from AI-generated fake evidence.
Emerging Judicial Thinking:
Need for forensic AI analysis and expert testimony to authenticate or debunk such evidence.
๐ KEY PRINCIPLES FOR ADMISSIBILITY OF AI-GENERATED EVIDENCE:
| Principle | Explanation |
|---|---|
| Reliability | Evidence must be accurate and trustworthy |
| Certification (65B compliance) | Electronic records require proper certification |
| Explainability | AI processes must be transparent and understandable |
| Expert Testimony | Experts needed to explain AI workings and outputs |
| Cross-examination | AI-generated evidence must be open to challenge |
| No blind reliance | Courts should not blindly trust AI results |
SUMMARY
AI-generated evidence is increasingly being introduced in courts, but it raises novel legal challenges. Courts in India have been cautious but open, requiring strong procedural safeguards like certification, expert interpretation, and cross-examination to ensure that AI evidence meets standards of reliability and fairness.
RELATED Blog
0 comments