PNB warns customers of AI voice cloning and deepfake fraud

Punjab National Bank has warned customers against increasingly sophisticated cyber frauds powered by artificial intelligence, including voice cloning, deepfake videos and hyper-personalised phishing messages designed to siphon funds.

The advisory from the state-run lender coincides with National Cybersecurity Awareness Month in October, as Indian banks modernise systems to protect depositors against emerging digital risks.

How criminals exploit digital habits

Scammers are moving beyond crude text lures to highly targeted social engineering that manipulates human emotion, according to the bank. Fraudsters employ cloned voices and synthetic video to impersonate family members, friends or authorities, fabricating sudden crises to rush victims into sending cash.

"Many cyber fraud cases today exploit human behaviour, particularly trust, fear, urgency and everyday digital habits," PNB Managing Director and Chief Executive Officer Ashok Chandra said in a message to customers. "Fraudsters can even imitate the voices or faces of your family members, friends or other trusted individuals and make false claims about emergencies to pressure you into transferring money immediately."

Chandra noted that fraudulent emails and text messages now closely mirror legitimate communications from lenders. By creating an artificial sense of panic, criminals push customers into transferring funds or entering credentials before checking the facts.

Strengthening India's digital security

The public alert reflects a broader push by domestic commercial lenders to fortify the country's financial architecture. Supported by regulatory standards set by the Reserve Bank of India, public and private sector banks are channelling capital into information technology infrastructure, threat detection systems and preparations against quantum computing risks.

Cybersecurity specialists emphasize that while technical infrastructure provides the primary shield, individual digital discipline closes critical gaps.

Srinivas L, Joint Managing Director and Joint Chief Executive Officer at 63SATS Cybertech, said fraudulent schemes consistently depend on artificial urgency, such as fake arrest calls or bogus warnings that an account will close within an hour.

"The technology has changed, but the playbook has not," Srinivas said. "It makes people panic, which pressurises them to pay."

Practical steps for account holders

Experts advise customers to pause whenever an incoming call or message demands instant money. Instead of replying to the incoming communication, recipients should hang up and phone the relative or institution back using contact details already saved in their address book.

Households can also establish an offline family code word. Because AI models cannot learn private verbal passwords kept within a household, a simple phrase confirms identity during distress calls.

Account holders are also encouraged to maintain basic digital hygiene: enabling two-factor authentication across banking apps and emails, choosing separate passwords, and establishing daily transaction limits on Unified Payments Interface transfers and debit cards so a single mistake cannot clear out an account.