Overview

• Artificial Intelligence (AI), while beneficial, has enabled sophisticated fraud.
• Deepfake-driven financial fraud is a growing threat; Pi-Labs projects ₹70,000 crore losses in India by 2025.

What Are Deepfakes?

• Synthetic media created via deep learning and neural networks.
• Produces realistic images, videos, and audio.
• McAfee survey: 70% cannot distinguish between real and cloned voices.

Risks & Modus Operandi

• Fraudsters impersonate executives or loved ones to authorise transactions or solicit funds.
• Challenges traditional KYC and phone verification; only 15 seconds of voice is needed for cloning.
• Video deepfakes simulate blinking, head movements, micro-expressions.
• Current usage: Video (46%), Images (32%), Audio (22%).

Impact

• Global 2024 losses: Average $450,000 per business (Regula).
• Long-term reputational damage; 72% of Indian firms rate it as the most severe fraud impact (KPMG).

Countermeasures

• AI & ML Detection: Analyse facial anomalies, lip-sync errors, voice timbre changes, and biometric cues.
• Blockchain Identity Verification: Immutable, encrypted, decentralised identities reduce cloning risk.
• Employee & Customer Training:
  • Recognise scam red flags.
  • Avoid reacting instantly to emergencies.
  • Use verification code words with loved ones.
  • Confirm media sources before action.