AI Has Upgraded the Fraudster

The Lift Line

Artificial intelligence has not invented fraud, it has industrialised it. From “digital arrest” scams that cost Indians Rs 1,935.51 crore in 2024 to deepfakes cloning Ratan Tata and N.R. Narayana Murthy, the fraudster has been upgraded. India’s response now spans new IT Rules, the DPDP framework, RBI safeguards and the 1930 helpline.

Why This Editorial Matters for Your Exam

This topic bridges cyber security, emerging technology, financial fraud and data governance, a rich source for GS3 Mains and current-affairs Prelims.

GS Paper 3: Awareness in the fields of IT, computers and cyber security; challenges to internal security through communication networks; role of media and social networking sites in internal security challenges; money-laundering and its prevention.

Prelims angle: IT (Intermediary Guidelines) Rules amendments, DPDP Rules 2025, Data Protection Board of India, I4C, Chakshu on Sanchar Saathi, MuleHunter.AI, IndiaAI Mission, relevant BNS sections.

Mains angle: Regulating deepfakes and synthetic content, the technology-versus-regulation race, and citizen-facing grievance mechanisms.

Background and Context

The signature scam of the moment is the “digital arrest,” where fraudsters impersonating police or agencies keep victims on a video call and coerce payments. No Indian law permits arrest via video call. The Prime Minister flagged it in Mann Ki Baat on 27 October 2024, offering the mantra “Ruko, Socho, Action Lo” (Stop, Think, Take Action), and urging citizens to report to the 1930 helpline or cybercrime.gov.in. Digital-arrest losses reached Rs 1,935.51 crore in 2024 (MHA).

The scale is large. Total cyber-fraud in 2025 is reported at about Rs 22,495 crore, with complaints up about 24 per cent to roughly 28.1 lakh (as reported by MHA and the Indian Cyber Crime Coordination Centre, I4C). I4C reports having saved Rs 8,189 crore and declined transactions worth Rs 9,055.27 crore till 31 December 2025, with 20,853 arrests. Deepfake investment scams have cloned Ratan Tata, Mukesh Ambani, N.R. Narayana Murthy and Sundar Pichai to endorse fake platforms.

The Core Argument / Issue

The regulatory response takes shape

MeitY notified amendments to the IT (Intermediary Guidelines) Rules on 10 February 2026 (in force 20 February 2026) to regulate “Synthetically Generated Information,” that is, deepfakes. These mandate labelling of AI-generated content, with audio required to carry a disclosure. Separately, the DPDP Rules 2025 were notified on 14 November 2025, though substantive obligations phase in by about May 2027, meaning they are not fully in force in mid-2026; the Data Protection Board of India (DPBI) is the adjudicator.

The financial-system defences

On the banking side, the Reserve Bank Innovation Hub (RBIH) built MuleHunter.AI, an AI tool to detect mule accounts, now live in about 26 banks. RBI’s April 2026 draft safeguards propose a one-hour delay on transfers above Rs 10,000, beneficiary whitelisting and two-factor authentication. The DoT’s Chakshu facility on the Sanchar Saathi portal lets citizens report suspected fraud communication before any money is lost.

The criminal-law toolkit

Under the Bharatiya Nyaya Sanhita 2023, Section 318 covers cheating, Section 319 cheating by personation, and Sections 111 and 112 organised crime, read with IT Act Sections 66C and 66D. The IndiaAI Mission (Cabinet, March 2024, Rs 10,371.92 crore) includes a Safe and Trusted AI pillar and an IndiaAI Safety Institute.

How to Think About This (Analytical Frame)

Frame this as an arms race between offence and defence, where AI lowers the cost of attack faster than it lowers the cost of defence. A single deepfake can be produced cheaply and scaled infinitely; detection and labelling must be built into every platform. The regulatory answer therefore works on three layers: prevention (labelling, whitelisting, transaction delays), detection (MuleHunter.AI, I4C interdiction) and redress (1930 helpline, Chakshu, criminal law). No single layer suffices; resilience comes from stacking them.

The Diagram in Words

Picture a funnel of defences a fraudulent transaction must pass through. At the top, awareness (Ruko, Socho, Action Lo) and content labelling filter out the obvious. Next, platform and telecom checks (Chakshu, intermediary due diligence) narrow the flow. Then banking safeguards (transaction delays, MuleHunter.AI, 2FA) catch the money in motion. Finally, if funds slip through, the 1930 helpline and I4C interdiction try to freeze them, and criminal law under the BNS closes the loop.

Way Forward

• Enforce deepfake labelling rigorously and build detection into platforms by default, not as an afterthought.
• Operationalise RBI’s transaction-delay and whitelisting safeguards while balancing user convenience.
• Expand citizen awareness of 1930, Chakshu and the “no arrest by video call” rule, especially among the elderly.
• Strengthen the IndiaAI Safety Institute as the standards body for trusted AI, and ensure DPDP enforcement begins on schedule.
• Improve cross-border cooperation, since much of the fraud infrastructure operates from outside India.

PYQ Linkage and Practice

Relevant PYQ threads: 2022 GS3 “What are the different elements of cyber security? Keeping in view the challenges in cyber security…”; 2018 GS3 on cyber-warfare; questions on data protection and privacy.

Practice question (15 marks, 250 words): “Artificial intelligence has industrialised financial fraud in India. Evaluate the adequacy of India’s layered regulatory and technological response.” Discuss.

Sources: The Indian Express

Source: AI Has Upgraded the Fraudster — Ujiyari.com | Free UPSC & State PCS Editorial Analysis