Artificial Intelligence Scams – New Cyber Attack

Artificial intelligence scams use AI-powered tools — voice cloning, deepfake video, and automated phishing — to impersonate trusted contacts and manipulate victims into transferring money or sharing sensitive information. These scams are more convincing than traditional fraud because they can replicate a real person’s voice or face in real time. In India, AI-enabled fraud is one of the fastest-growing cybercrime categories, with victims often unaware they are speaking to a machine or a criminal.

How Do AI Scams Work?

AI scams exploit the trust we place in familiar voices and faces. The two primary techniques are:

1. AI Voice Cloning

Cybercriminals use AI voice cloning tools to replicate the voice of someone the victim knows — a family member, friend, or boss. The clone is created from just a few seconds of audio gathered from public videos, social media voice notes, or call recordings.

The attack proceeds as follows: the fraudster calls the victim using a spoofed number (see: caller ID spoofing) that appears to be the real person. They speak in the cloned voice, create a sense of urgency (“I’ve had an accident”, “I need money urgently”), and ask the victim to transfer funds immediately. The victim, believing they are speaking to their loved one, complies.

This attack is particularly dangerous because every voice is a unique biometric identifier. The prevalence of voice data on social media — audio clips, video calls, voice notes — means virtually anyone with a public online presence can be cloned.

2. Deepfake Video Calls

AI deepfake technology replaces the face and voice of the attacker in a live video call with that of a trusted person. The victim sees and hears someone they recognize — but the person is not real. This is being used in CEO fraud attacks on businesses and in romance scam operations.

Deepfake calls are also used for online sextortion — where criminals create fake video calls appearing to be a romantic interest and capture the victim in compromising positions for blackmail.

3. AI-Generated Phishing

Tools like WormGPT generate personalized, grammatically perfect phishing emails at scale. Unlike traditional phishing emails that are easily identifiable by poor grammar or generic content, AI-generated phishing emails reference specific details about the target — their employer, recent activities, or contacts — making them extremely convincing.

4. AI Chatbots for Customer Care Fraud

AI-powered chatbots impersonate the customer service portals of banks, payment apps, and government agencies. Users searching for “bank customer care number” or “government helpline” on Google may land on scam websites where an AI chatbot collects their credentials and OTPs in real time, forwarding them to attackers. This is a variant of customer care fraud.

Why AI Voice Cloning Is Especially Dangerous in India

India has one of the highest rates of voice and video sharing on social media globally. Voice notes on WhatsApp, short video clips on Instagram Reels, and video calls on Zoom all provide the raw audio data criminals need to clone a voice. A convincing clone can be created from as little as 3–5 seconds of audio using commercially available AI tools.

How to Protect Yourself from AI Scams

Establish a Code Word with Trusted Contacts

Create a private code word with family members and close friends. If you receive an urgent call asking for money, ask the caller for the code word before taking any action. No legitimate emergency prevents someone from providing a pre-agreed code word.

Pause Before Acting on Any Urgent Request

AI scams rely on urgency to prevent the victim from thinking critically. Never transfer money, share OTPs, or provide personal data based on a single phone call or video call — regardless of how authentic it sounds. Hang up and call the person back on their known, saved number.

Verify the Identity of the Caller

Ask cross-questions that only the real person would know — specific shared memories, mutual contacts, or details from a recent conversation that could not have been sourced from public profiles.

Do Not Trust Caller ID

Caller ID can be spoofed using VoIP technology to show any number. A call appearing to come from your mother, your bank, or even India’s cybercrime helpline does not guarantee it is genuine. Always verify using a separate call on the person’s known number.

Limit Voice and Video Data Exposure

The less audio and video data you share publicly, the harder you are to clone. Review your social media privacy settings and limit who can access your voice notes, videos, and reels.

Enable Two-Factor Authentication

Even if an AI scam obtains your password, two-factor authentication prevents account takeover without the second factor.

What to Do If You Are Targeted by an AI Scam

1. Do not transfer money. If you already have, call your bank immediately to attempt a transaction freeze.
2. Collect all evidence — call recordings, screenshots, transaction references.
3. File a complaint at cybercrime.gov.in or call the National Cyber Crime Helpline: 1930.
4. For professional investigation and evidence analysis, contact a cyber expert in India.