Banking in the AI age needs more than speed

ARTIFICIAL intelligence (AI) has made scams more convincing, quicker and harder to detect.

A fake video can make it look as if a public figure is endorsing an investment, and a fake banking link can look polished enough to fool even a careful person.

That is the truth of modern fraud. It is no longer enough to tell Malaysians to “be careful” because even careful people are getting cheated.

Syndicates are using AI tools such as deepfake videos, automated chatbots and voice impersonation to trick victims.

With most of the financial world now online, crooks don’t need to rob banks these days. They just need to panic a victim, steal a password, take over a phone and move money into a mule account.

This is why banks and financial institutions must treat fraud prevention as part of their products, not as an aside on their portal.

Malaysia has already taken useful steps. Banks no longer rely only on SMS one-time codes for some transactions while cooling-off periods, single-device approvals, 24-hour fraud hotlines and kill-switch functions have become part of the response.

These are important, but the next stage must be stronger, simpler and more human.

Banks should slow down suspicious transactions before the money disappears. Not every large transfer is fraudulent, but a sudden transfer to a new account at odd hours after a password change from a newly registered device should raise a red flag.

The customer may find it slightly inconvenient, but a short delay is better than a lifetime of regret.

Banks should also design protection for ordinary users, not only confident users. Many ­elderly customers, small traders, retirees and rural customers may not fully understand app permissions, fake links, remote access software or QR-code fraud.

For them, banks should offer a safer banking mode. This could include lower daily transfer limits, stricter approval for new recipients of funds, branch confirmation for high-risk changes, and the option to block online overseas transfers unless requested.

Banks must also make the kill switch easier to find and use. In a panic, no victim should have to search through five menus to freeze an account.

Financial institutions should also improve how they deal with victims after a scam. Too often, victims are made to feel foolish.

This is wrong because shame helps scammers by delaying reporting. Banks should train frontline staff to respond quickly and respectfully because the first few minutes matter.

The faster an account is frozen, the better the chance of stopping the money from disappearing.

Banks, telcos, police and regulators must also share information faster. Scam accounts should not be allowed to move from one bank to another as if nothing happened.

Mule accounts must be detected earlier, closed faster and investigated properly. If the system knows an account is receiving suspicious funds from multiple victims, action should not wait until the damage becomes huge.

When banks build the road, they must also build the guardrails. The aim is not to make banking difficult, but to make it safer for everyone.

At the same time, customers also have responsibilities. They should not share passwords, download unknown apps, click suspicious links or believe promises of guaranteed or high returns.

AI has changed how scams work. Banks must now change the way they protect customers. Digital banking cannot be judged only by how fast money moves. It must also be judged by how well it protects people.