How to prevent ATM frauds using Artificial Intelligence?

With the increase in digital payments, ATM banking scams have also increased. Scammers have been using various techniques to cheat banks and customers via ATM skimming. And, reports of such cases have increased in recent months.

Hemant Warier
Published in
4 min readAug 3, 2020


A pictorial representation of a fraudster stealing ATM card details

One of the most common types of ATM fraud that exists is ATM cloning or ATM skimming. ATM skimming is a process where thieves place a device on the face of an ATM, which looks like a part of the machine.

The criminals often hide a small pinhole camera in a brochure holder near the ATM. Then they extract the target victim’s ATM PIN. The criminals hide the camera from the view. When an unknown victim uses their card, the criminals capture the card details and the PIN number.

Global Losses

Recently in Alaska, the local police caught a man who has a history of International skimming. Now he is being charged for possession of a counterfeit access device in U.S. District Court in Alaska. He’s used over 12 different aliases. He admits of skimming for more than $200 thousand only in Utah.

The global losses due to the activities of ATM scammers are running into billions of dollars. The Nilson Report expects that worldwide loss due to ATM fraud will rise above $40 billion by the year 2027. By the end of the year 2020, total losses would be around $33.5 billion.

A graph depicting the Nilson report, which shows the increase in worldwide losses due to ATM Skimming & other ATM frauds.
Gross fraud losses reached $27.85 billion in 2018, up 16.2% from $23.97 billion in 2017. However, fraud losses per $100 of total sales declined to 6.86¢ from 6.95¢ the prior year.

According to the article on PR Newswire, more than 55% of all fraud-related complaints list a method of initial contact. Out of which 77% are being done by phone, 3% are being carried by mail, while 8% by emails. Most of the emails are a part of a phishing attack that comes from unknown sources. Also, 65% of the time, credit card fraud results either in a direct or indirect fiscal loss for an individual.

Actions taken by Banks

Recently the ATM fraud reports have increased. Many banks have been sending warning messages to their customers.

State Bank of India has tweeted preventive measures for customers from ATM fraud.

The State Bank of India (SBI) has said that it will refund the stolen money to the ATM fraud affected customers.

Spain’s CaixaBank is using facial recognition technology on 20 ATMs since early 2019. The biometric authentication capability lets consumers withdraw cash without a PIN. According to ComputerWeekly, the bank is adding 100 more facial recognition machines.

Agriculture Bank of China, one of the ‘Big four’ in the country, uses facial recognition based withdrawals in their ATMs since 2017. This has helped the bank to:

  • Reduce the use of counterfeit cards
  • Improve customer experience as customers don’t need their ATM cards to withdraw money.

How can FaceQuest help?

FaceQuest offers 1:1 Face Match in which a user can upload two different photographs and verify if they are of the same person. Banks can integrate FaceQuest’s facial recognition AI with ATM servers to add an extra layer of security. The banks can store photographs of their account holders in FaceQuest’s encrypted server.

Photograph of the account holder stored as a reference face
Photograph of the account holder stored as a reference face


  • The bank’s ATM machine will click the photograph of the user and match it with the account holder.
Photograph of the ATM user clicked via ATM machine
Photograph of the ATM user clicked via ATM machine
  • FaceQuest will be able to verify the match of both the photographs in less than 2 seconds.
  • A successful match of the photograph will authenticate the transaction.
FaceQuest shows the completed match result with correct detection of face
FaceQuest shows the completed match result with correct detection of face
  • In case the facial match is not confirmed, FaceQuest will trigger the alert to the bank. The banks can then send an OTP to the customer via email/phone.

To know more about FaceQuest’s Facial Recognition as a Service, schedule a call with the team.