Hyderabad: Amid the rising menace of mule accounts — used by cybercriminals to divert illicit funds — experts stressed the need for banks to be more alert in detecting such frauds.
Speaking on Day 2 of ‘SHIELD 2025,’ a cybersecurity conclave organised by the Telangana Cyber Security Bureau (TGCSB) and the Society for Cyberabad Security Council (SCSC) in association with The Times of India, Rajesh Bansal, CEO of the Reserve Bank Innovation Hub (RBIH), said, “We did a study with 10 banks and found that eight of them use conventional and rule-based systems to identify mule accounts to date. Banks also tend to question the return on investment for solving these mule account problems.”
Nilesh Deshpande, head of the financial crime prevention group at ICICI, echoed the same concern, “Banks do not educate customers when they come to open their accounts. They aren’t explicitly told that they cannot lend or rent their accounts to anyone and that there are legal consequences for doing so.”
208 bn transactions in 2024
But then there are challenges that need to be addressed, said experts. The scale of digital transactions in India, for instance, has made detecting fraudulent activities increasingly difficult.
“It’s surprising to many, but the number of digital transactions done in India in 2024 was 208 billion. In such scenarios, it’s so hard to trace which transaction is good or bad,” said Krishna Sastry P, partner, cyber security at Ernst & Young. He added, “From rickshaw pullers to vegetable sellers, common people are being reeled in. Approximately 0.7% of all accounts in India are involved in mule transactions.
Deshpande added, “The major issue at hand is how we identify mule accounts from normal accounts. Mule accounts are like sleeper cells — unless activated, we can’t identify them. We are instead relying on transaction behaviour to ascertain what’s what.”
AI/ML-driven solutions
Offering solutions, the panel affirmed that newer systems involving AI and ML, along with greater collaboration with banks, must be adopted to address the issue.
Bansal said, “In 2023, we started with UPI frauds but realised we needed to do something foundational. Currently, I’m happy to tell you that the govt and RBI recently launched MuleHunterAI. It’s deployed in three or five banks, with more banks approaching us. It can take even 20 minutes to do what previously took four weeks.”
MuleHunterAI is an AI/ML-based solution by RBIH that helps identify mule accounts. It can analyse transaction and account-related datasets to predict mule accounts with higher accuracy and greater speed.
V Radha, associate professor at the Institute for Development and Research in Banking Technology (IDRBT), emphasised the need for advanced technology. “Technology has dramatically improved over the years. I don’t think rule-based systems will work anymore. AI will help in detecting mule accounts and can solve false positive cases that pop up,” she stated.
Deshpande also pointed out a geographical pattern: “We’re also seeing a concentration of mule accounts in urban areas. So, at the branch level, while engaging with customers, we need to educate them that they cannot give away their bank details to anyone for commissions or to make a quick buck.”
The experts advocated for a multi-pronged approach to combat mule accounts through customer education. As Deshpande summarised, “This is a total tech war. AI and ML will learn from every transaction. It’s the fraudsters who will teach us how to catch fraudsters.”
Speaking on Day 2 of ‘SHIELD 2025,’ a cybersecurity conclave organised by the Telangana Cyber Security Bureau (TGCSB) and the Society for Cyberabad Security Council (SCSC) in association with The Times of India, Rajesh Bansal, CEO of the Reserve Bank Innovation Hub (RBIH), said, “We did a study with 10 banks and found that eight of them use conventional and rule-based systems to identify mule accounts to date. Banks also tend to question the return on investment for solving these mule account problems.”
Nilesh Deshpande, head of the financial crime prevention group at ICICI, echoed the same concern, “Banks do not educate customers when they come to open their accounts. They aren’t explicitly told that they cannot lend or rent their accounts to anyone and that there are legal consequences for doing so.”
208 bn transactions in 2024
But then there are challenges that need to be addressed, said experts. The scale of digital transactions in India, for instance, has made detecting fraudulent activities increasingly difficult.
“It’s surprising to many, but the number of digital transactions done in India in 2024 was 208 billion. In such scenarios, it’s so hard to trace which transaction is good or bad,” said Krishna Sastry P, partner, cyber security at Ernst & Young. He added, “From rickshaw pullers to vegetable sellers, common people are being reeled in. Approximately 0.7% of all accounts in India are involved in mule transactions.
Deshpande added, “The major issue at hand is how we identify mule accounts from normal accounts. Mule accounts are like sleeper cells — unless activated, we can’t identify them. We are instead relying on transaction behaviour to ascertain what’s what.”
AI/ML-driven solutions
Offering solutions, the panel affirmed that newer systems involving AI and ML, along with greater collaboration with banks, must be adopted to address the issue.
Bansal said, “In 2023, we started with UPI frauds but realised we needed to do something foundational. Currently, I’m happy to tell you that the govt and RBI recently launched MuleHunterAI. It’s deployed in three or five banks, with more banks approaching us. It can take even 20 minutes to do what previously took four weeks.”
MuleHunterAI is an AI/ML-based solution by RBIH that helps identify mule accounts. It can analyse transaction and account-related datasets to predict mule accounts with higher accuracy and greater speed.
V Radha, associate professor at the Institute for Development and Research in Banking Technology (IDRBT), emphasised the need for advanced technology. “Technology has dramatically improved over the years. I don’t think rule-based systems will work anymore. AI will help in detecting mule accounts and can solve false positive cases that pop up,” she stated.
Deshpande also pointed out a geographical pattern: “We’re also seeing a concentration of mule accounts in urban areas. So, at the branch level, while engaging with customers, we need to educate them that they cannot give away their bank details to anyone for commissions or to make a quick buck.”
The experts advocated for a multi-pronged approach to combat mule accounts through customer education. As Deshpande summarised, “This is a total tech war. AI and ML will learn from every transaction. It’s the fraudsters who will teach us how to catch fraudsters.”
