⚠️ Emerging AI Risks
🎭
Deepfakes
Synthetic media for fraud/defamation
🗣️
Voice Cloning
AI-generated voice for impersonation
📝
AI-Generated Text
Fake news, phishing content
🎨
Synthetic Images
Non-consensual intimate images
⚖️
Algorithmic Bias
Discriminatory AI decisions
🔍
Surveillance AI
Mass facial recognition
📋 Legal Framework for Deepfakes
⚠️ Current Gap: No Specific Deepfake Law in India
India lacks a dedicated deepfake law. Cases are prosecuted under existing provisions with limited effectiveness.
Applicable Provisions
| Law | Section | Application to Deepfakes |
|---|---|---|
| IT Act | 66C | Identity theft using AI-generated content |
| IT Act | 66D | Cheating by personation using computer resource |
| IT Act | 66E | Violation of privacy (synthetic intimate images) |
| IT Act | 67/67A | Publishing obscene/sexually explicit material |
| IPC/BNS | 499-500/356 | Defamation through deepfakes |
| IPC/BNS | 509/79 | Outraging modesty of woman |
| Copyright Act | 51-63 | Infringement using AI-generated copies |
📌 IT Rules 2021 — Intermediary Obligations
Rule 3(1)(b)(vii): Intermediaries must remove content that impersonates another person. Rule 3(2)(b): 24-hour takedown for content depicting private parts of individuals.
⚖️ Algorithmic Harm & Discrimination
AI systems can perpetuate or amplify discrimination in lending, hiring, insurance, and criminal justice.
Types of Algorithmic Harm
- Discriminatory Outcomes: AI denying loans/jobs based on protected characteristics
- Opacity/Black Box: Inability to explain decisions affecting rights
- Manipulation: Algorithmic nudging affecting consumer choices
- Surveillance: Pervasive tracking and profiling
✅ Legal Remedies for Algorithmic Harm
- Article 14: Constitutional challenge for arbitrary discrimination
- Consumer Protection Act: Unfair trade practice, deficiency in service
- DPDPA: Right to information about automated decision-making
- Sectoral Regulations: RBI, SEBI, IRDAI norms on algorithmic fairness
🎯 MeitY Advisory on AI (March 2024)
📌 Key Requirements for AI Platforms
- Permission from Government before deploying untested/unreliable AI
- Label AI-generated content to identify synthetic media
- Prevent generation of unlawful content
- Ensure AI doesn't permit bias/discrimination
📝 Part 12.5 Quiz
Q1: Section for identity theft under IT Act:
Q2: Section for cheating by personation:
Q3: IT Rules 2021 takedown for intimate images:
Q4: Privacy violation (synthetic images) under:
Q5: MeitY AI Advisory issued in:
Q6: Algorithmic discrimination can be challenged under:
Q7: Publishing obscene AI content falls under:
Q8: India has specific deepfake law:
Q9: IT Rules require labeling of:
Q10: Voice cloning for fraud can be prosecuted under: