By Shashank Didmishe and Shubhra Tandon
At a time banks are adapting to technology at a rapid pace and making strides in improving speed of transactions, a clutch of technology officials believe that artificial intelligence (AI) and machine learning (ML) have a huge role to play when it comes to fraud detection and enabling banks to fulfill their compliance needs. However, as of now banks are not equipped with the necessary know-how to deploy such solutions, they added.
At present, the treatment given by the banking sector to frauds is reactive such that any action taken to provide relief to the customer is done after the fraud has already been committed. Banks can benefit from patterns that emerge with the use of AI and ML to prevent frauds even before they happen.
“What AI is doing now is that all the patterns that are emerging, we can train the models in a manner that they can forecast and predict with a level of accuracy that we would like to have. Since we don’t produce this technology at banks ourselves, we are depending on technology providers,” Nitin Chugh, deputy managing director and head of digital banking at State Bank of India said while talking at the financialexpress.com’s Modern BFSI Summit.
In order to prevent frauds, banks need to have robust mechanisms as with the rise of digital banking, multiple transactions are going on through multiple channels which need to be monitored. The task of detecting frauds on a preventive basis is more suited for a machine and untenable for humans. The question is not whether we require AL and ML or not, as with thousands of transactions happening digitally per second, banks need to essentially to deploy these mechanisms without fail.
“As digital expands and explodes, frauds also tend to expand at a certain speed need to be addressed. AI-based models can aid in fraud detection. AI can help in velocity of detection,”Jagdish Narayanan, senior vice president, Jio Financial Services.
Emphasising on the volume of compliance burden on banks, Akhil Handa, chief digital officer, Bank of Baroda said that the lender files around 600 regulatory filing per month, the application of technology. There is a lot to be done and the application of AI and ML to compliance system is evolving, he added. Additionally, the framework of the system is so rigid that the regulator needs to ensure that the banks are aware and safe, Bijith Bhaskar, national head of digital banking at ICICI Bank said.
There is a perception against AI and ML that it makes human workforce obsolete; however, that argument against technology is not valid, according to Handa. With AI the efficiency of the entity increases, which although takes away the workload of certain section of the workforce, it provides scope to create new roles within the same space by upskilling the existing workforce.
“The other thing is that banks have traditionally used customer-initiated data. So there is always a scope of error. Most of the models that run on AI use machine data,” Chugh said. Machine data is rather pure and it does not have those sets of errors that humans might make while inputting data. “So as we move to more digital transactions, move to more machine-led interactions on mobile or otherwise, the data purity will also improve and that will help us in defining the model,” he added.