0:13 Namaskar viewers. Hello and welcome to Sensor TV. I am Tina Ja. You're watching Perspective. As India's digital economy expands at an unprecedented pace, so do the challenges that come along with it. From online banking frauds to sophisticated cyber scams, cyber criminals are constantly finding new ways to exploit the vulnerabilities in the digital ecosystem. One of the biggest concerns today is the growing misuse of mule accounts, bank accounts which are used to channel and hide fraudulent transactions. So in a significant step towards building a more cyber seccure bhat the Indian cyber crime coordination center under the ministry of home affairs and the reserve bank innovation hub have joined hands to strengthen the fight against cyber enabled financial frauds. The two
0:58 institutions have signed a memorandum of understanding aimed at boosting fraud risk intelligence sharing, enhancing analyt analytical support and improving operational coordination across the banking and digital payments ecosystem. Now, this collaboration seeks to harness the power of artificial intelligence through platforms like mulehunter.ai, enabling faster detection of suspicious accounts and strengthening proactive fraud prevention systems. Union Home Minister Amit Sha has described this initiative as a next generation shield against cyber crime underlining the government's commitment to securing India's digital future. So how will this partnership work? How can AI transform fraud detection and what does it mean
1:44 for citizens, you and I, the banks and the country's rapidly growing digital payments ecosystem. To discuss some of these important aspects and more, I'm joined by distinguished experts. It gives me immense pleasure to welcome on this edition of perspective Mr. Nishant Kumar. He's director Indian cyber crime coordination center ministry of home affairs. Mr. Kumar, thank you for taking our time joining us on Sunset TV. Welcome to the program. With me in the studio is Dr. Utpal Chakrabati. He is co-founder and chief technology officer Indic AI. Dr. Chakrabati, thank you for joining us in the studio today. And joining us virtually is uh Dr. Prashant Mali. He's an advocate and also a cyber and data protection expert. Dr. Kumali, thank you for joining us on the program. Welcome to Sunset TV. Uh Mr. Kumar, so let me begin the program today with you.
2:30 Uh first to understand the context of how this uh partnership uh happened. What prompted the Indian Cyber Crime Coordination Center and the Reserve Bank Innovation Hub to come together for this important collaboration from a policy perspective? if you could explain to us and our viewers how important is thisou in India's fight against cyber enabled financial frauds. >> Yeah. Uh very good evening to all the viewers and uh to the sun TV. Uh basically uh see uh as you know that the Indian cyber crime coordination center which we normally call as popularly as I4 under MHA we are the model agency to you know fight against cyber crime in the country. And uh when you talk about the type of cyber crimes, the financial
3:17 crime are the main you know what we deal with the main challenge that comes because uh the financial transactions in India as you know is happening at a very you know fast pace because of the rapid digital dig digitalization that has happened. >> That's right. And in that case, yeah, in that case, if you see, it's not something new that we have collaborated with RBI or banks because when we talk about financial cyber crime, then of course we have to, you know, deal very closely work with the banks and the financial institutions. And of course, since RBI is the you know, you can say the regulatory body whenever we want to implement a policy with all the banks, we used to channelize it with RBI. And in fact as I I'd like to tell you that with the instructions of honorable HM
4:03 the I4C and RBI are having very you know periodical meetings multiple times in a year to uh to you know to discuss what type of challenges are coming in the financial ecosystems and then uh how to counter them what we have done in the past we to take a look on that how much those you know actions have worked and what all improvements are needed. So based on all these things uh see we come came across this uh because this fact that mu accounts as you all know and those who don't know that m accounts are those fraudent accounts which are opened by some you know fraudsters or some people who are knowingly or unknowingly creating these opening these accounts and these accounts are used to carry the
4:50 fraudulent money. So when when we talk about cyber fraud the money when it debits from a victim's account it goes to multiple layers which we normally call layer zero layer 1 layer two layer three and then it goes to layer and from where it is it is taken out withdrawn or maybe converted into a cryptocurrency and then sent to some abroad or maybe taken out from a an ATM. So in that case what happens that when the investigation takes place we have to deal with these multiple fraudulent news accounts. So if we want to curtail or curb these cyber crimes, if we bring this mule account to a you know to a you can say we want we can bring this mu account or restrict the opening of this mu account the channel the path with which this
5:36 fraudulent amount is being rooted that stopped and therefore there was a big challenge with all the banks because we have been giving data and uh you know inputs to all the banks that there are multiple new accounts which are being opened in the banks because of certain you know uh the courts are using certain you know gimmicks to get those accounts open. Maybe a person himself goes and you know rents his account or maybe unknowingly his account is being used by somebody. So how banks can you know work and find out these new accounts stop the opening of these new accounts and then if new accounts are opened how they can you know close them or how they can restrict them. So uh this is this challenge was there in front of banks for long
6:21 >> right. So this was under discussion with banks as well as RBI and between MHA, IOT and based on that since the AI you know the AI technology has now come for and everybody's using even the forers are using these AI accounts to create deep and for the you know common citizen or common man in that case we uh the RBI innovation hub which we call RBI it came up with this plan that since multiple banks are using their own technology to you know find out and detect and then to close or restrict these new accounts. So RBI centrally it worked on some technology using AI so that using certain parameters or certain you know the character of the banking
7:09 account like account is openly open very new and then and very suddenly it starts getting lot of money which is coming and going from those accounts. So since there are lacks of accounts, flows of accounts that a bank has to handle. So AI could detect can detect these anomalies that are happening with those you know new account those accounts which can be uh you know tagged as new account. >> Right. Right. >> Yeah. So automatically identifying and alerting the banks so that they can work and uh take action on those accounts. So with this view uh or with this you know idea this uh idea came and then RBI IH and I4 because we have a certain data we have a big suspect registry where all the fraud type of accounts and also when
7:54 the police and state police also using multiple account and then his linkages are found. >> Sure. Sure. So Mr. Kumar we we'll uh try and understand about uh the uh registry the I4 C suspect registry how it's going to improve uh uh you know the effectiveness of the AIdriven fraud detection models but what we understood from your uh intervention is that the focus is on the mule accounts and it was interesting to hear how uh given the challenge that a country like India was facing uh this AIdriven system will be able to now identify uh the suspicious and the mule accounts at a much faster pace. Uh that was the policy perspective that you shared with us. Uh I'll come
8:39 now to you Dr. Mali from uh uh now for a legal perspective since you're also an advocate. How important according to you is this partnership of course to begin with between the I4C and the RBI. Uh do you think it's going to mark a significant shift uh towards say a more coordinated a more proactive cyber crime prevention framework something that was much needed in a country like ours? >> As a practicing lawyer let me first congratulate uh I4C for taking this step of bringing predictive policing instead of reactive policing until till date what we were doing uh using this 100.ai AI and the database. Sharing the database with this AI gives us hell lot
9:25 of you know algorithmic intelligence for financial cyber warfare and this is going to be the best thing to happen to stop cyber crime cyber crime related frauds because I have faced cases where I have seen villages the whole village was into uh you know uh accounts which they were holding mule accounts and they were lured for 10% % or 5% of money which comes to their account and these people were told that this is some kind of business which has come up and people were just lured. Sometimes relatives were lured into it and this has become a whole big pain because police were not able to reach to the actual cyber criminals because they ended up in
10:10 catching this mule accounts. If this mule accounts they are the blood vessels of cyber crime you know if we attack on those blood vessels then the economy of cyber crime gets automatically destroyed. So now you know if we know that this meal account you know exist and they if they are targeted then the place where the money was taken by the cyber criminals itself is attacked and if itself is stopped then they may have to find something else to do but at least this is the right step I feel uh for many many cyber crime victims because every kind of cyber crime whether it is digital arrest whether Whether it is stock market eye fraud or
10:56 whether it is uh you know making profiles and then looting or fishing kind of frauds any kind of cyber frauds mule accounts was the important cyber weapon in this whole cyber game of cyber warfare and if this is attacked I think this would be a pre preactive you know proactive kind of uh step which uh we are do going to take for controlling cyber crime in India. >> That's right. Dr. Shakrabati, what are your thoughts? Do you think it's it's that much needed step that India's fast growing digital ecosystem needed and and it's towards at least institutionalizing uh cyber fraud prevention. So a lot was spoken about it now that step has been
11:41 taken. Yeah, I think uh this is a major step uh that has been uh taken and uh this is required at least at the in the ease of AI uh because what happens like uh so this is a kind of laundering. So we have seen the laundering cases uh in the banking uh sector but this is a special type of uh laundering using the mule accounts. Now what AI can do is AI can uh see the patterns uh how the accounts are created, how the transactions have been made and where the money is flowing whether within the country or cross border and how the accounts are been getting closed. So these patterns uh AI can uh automatically detect and any future
12:27 transactions similar kind of transactions happening it can identify that pattern and can highlight that these are uh the accounts which are actually doing this fraudulent activities. So this is uh the major contribution that AI can bring there because otherwise we may write some rules traditional uh but as you know this uh landscape uh the cyber crime landscape is uh changing in a very fast space and this cyber gimnals especially the the financial frauds they are coming up with newer and newer techniques. That's right >> and to only way to deal with these techniques is basically to put a uh a system which is adaptive. Now AI is adaptive system and using AI you can actually identify this in both the ways
13:13 like in a supervised way and also in unsupervised way. Supervised way in the sense like the model or the machine AI machine is trained with similar kind of scenarios and it can identify those and there is also unsupervised way which the other panelist has uh just mentioned is basically it can do the anomaly detection. So any kind of anomaly that means which is not usual okay because all the transactions that are happening in the banking ecosystem has a pattern and if there is a uh is a is a pattern which is not aligned with the normal pattern that we see then AI can easily highlight and detect those and I think this is going to be a great help uh for uh the cyber analyst who will be sitting
13:59 at the RBI and also at the national uh this thing in the this thing >> in fact uh uh Mr. Mr. Kumar did mention about the suspect registry. So let me go across straight to him to understand more about the suspect registry which has been developed by I4C how it works and how it it's going to work uh uh or support the banks and the financial institutions especially given that one big challenge in the area of cyber fraud Mr. Kumar is the speed. So money moves within seconds. How is it going to work if uh and how are you going to track all of that is something that I believe not just me but my viewers will also be quite curious to know about. >> Yeah, very true. So just like one of our panelists mentioned that proactive policing. So basically what happens in cyber crime normally when a crime
14:46 happens means the person is losing his money and then he's reporting it and then action is being taken and in that case as you rightly said that time holds the essence like if the complaint is made you know after few hours of the happening of the loss of the money it's very difficult to get it you know because we have to stop it in the financial ecosystem. So uh we are thinking about various proactive measures. One of those measures is the suspect registry just like a mule hunter we are talking about. So in the split registry what happens that we are trying normally what happens we are getting multiple type of calls and we most of the time we understand this is a fake call this is a fraudulent call and what we do we disconnect the call and then we block the number but what if a person is
15:32 reporting that number that this is a number which belong to a suspect we call it a suspect because I feel it's a coordinate number so when a citizen reports those you know coordinate numbers and similarly when you know a amount is getting channelized to a account which we suspect that it can be a m account or whenever a criminal is arrested and then the accounts in which he has warmed upon. So we take all those accounts as a mule account because see if we know one real account it may be related to various other accounts where the money is going through it. Those accounts may or may not have been reported but we can understand that those accounts which they are related to a proven mule account they also can be mu accounts. So in this way we are getting multiple type of you know
16:19 you can say early you know messages that these type of phone numbers these you know accounts or these maybe some uh you can say some WhatsApp links some WhatsApp groups some social media links some websites which all can be a you know tool of cyber crime which we suspect. So what we ask citizens, we ask banks and various other stakeholders to report those suspected you know accounts those numbers to IOC and when those come to our place we collect it in a repository and similarly there are banks and other entities also who are collecting these type of you know suspected entities >> and then we you know we collap in in the back end and then we create a bigger
17:04 suspect registry which helps us in taking preventive action against those suspects. What happens when we we detect something as a we you know categorize something as a suspect then we take further action to identify and you know finalize that yes they are actually a you know fraudulent account or a fraudulent number and in that case we take action of blocking. You must have heard that these number of you know pin numbers have been blocked or these many numbers of phone IMI numbers have been blocked or these many accounts have been you know deactivated by bank. So in all those cases they are all the you know out output which are coming out of the this intelligence is coming out of the suspect registry. So we ask citizens and everybody and banks and all the other
17:50 entities to report those suspected accounts other entity other you know mobile numbers or website internet you know artifacts to us and then we take it in the suspect registry. we cross it and identify it with various other you know you can say parameters and if we finalize that these are actually problems. So these can be anything. This can be bank account numbers. This can be UPI ids. These can be email ids. These can be social media handles, website URLs, IMEI. And then we once it is flagged as a suspect, we take action and then we you know take action to block, deactivate or you know bring them down so that they cannot act further as a tool to commit cyber crime.
18:36 >> That's right. So intelligence and analytics these are going to be very very crucial in making this initiative work. Since you spoke about collaboration, Mr. Kumar, a related question is that uh uh cyber fraud investigations often involve multiple stakeholders. So, I'm going to lead that question to collaboration between various institutions. So, the stakeholders include the banks, the payment gateways, uh not to forget also the telecom operators, but most importantly the police agencies, uh the law enforcement agencies. Uh is it safe to say or is it is it time that we say this initiative can strengthen the institutional coordination in tackling such crimes uh at speed at scale. >> Right. Very true. You also missed one social media and the media is they are
19:23 also a very prominent stakeholders because social media nowadays is you know omniresent and they act as a many time most of the times I see that they act as a platform which are used to commit cyber crimes because you know there are various social media ads which are you know prevalent in nature which are using defect and then people are getting because how a person comes to a you know investment scam uh in the trap of investment STEM or some other digital because all those are coming through these social media ads. So in that case we are cooperating coordinating with social media intermediaries like Meta, Google and others and they are also giving a good cooperation to us. In fact uh if you would like to know uh we have developed one all more application which
20:10 is called Sahob which means cooperation in fact >> and there all these stakeholders have been onboarded. So what happens whenever we find any these type of suppose there is one part which is called suspect there's other part where we know that this link is a fraudulent link because it has been reported multiple times it has been flagged by the police so through that your portal we are sending those fraudulent links those fraudulent or any other number or any other thing to the other stakeholders and they are immediately taking it down within a stipulated time. So this has been applauded by the honorable court of India and this is a very proving to be a very good tool so that we are a with a very limited time we are bringing able to be able to bring down those you know internet based uh you know uh
20:57 information or traps which are used to commit cyber crimes and which can lure people to fall in the >> uh what you've explained uh you know it indicates that public trust is very very important for the success of digital India and it's initiatives like these that you mentioned uh will go a a significant mile to uh you know strengthen the citizens confidence in the digital ecosystems that we are strengthening day by day. Uh Dr. Mali, so now as artificial intelligence becomes central to fraud detection. Uh what are the legal and the ethical safeguards that become necessary at this point in time? uh how do you think the various authorities involved in in its implementation will now balance effective fraud monitoring uh with the
21:44 concerns around privacy and data protection? Yeah, this is a band-aid question. Like when we look at the real rosy side of this, we also need to focus on the huge amount of data being shared by uh you know whether it is suspect data or banking data because that all gets classified as personal data and when when a personal data comes you need to follow DPDP and uh since right now what we need the most because we are using AI at this great level the AI ecosystem needs to establish trust trust in the whole system and to establish this trust we need AI governance related rules or AI laws and when we have this AI laws in
22:30 place because until unless whatever good features uh I4C with good intention and RBI is promoting if we don't have the framework of law it can always be challenged in the courts so to avoid this we need to fix this in a framework of law uh using usage of because the intention is good. Intention is to prevent crime. So why not you know give give it a framework of law and frame some rules because when it suppose it gets appealed in courts of law or any constitutional PL is been put then these rules which are published rules on AI based monitoring and with safeguards in place we know it is easy to convince to the courts that constitutionally
23:16 using the constitution is not being you know trampled upon the constitutional rights are not been trampled upon of privacy or uh uh you know secrecy rights are not trampled upon but it is only for uh you know giving justice to cyber crime victims and I think honorable justices of supreme court and high courts are very well aware that cyber crime has become a big menace and this you know hunter.ai AI would act as a digital constable who is watching the banking ecosystem 24 by7. Now this gives him other power you know apart from pattern recogn recognization and etc. 24 by7 watching banking ecosystem which gives us edge to catch
24:03 the frauds because many times it is also see this frauds happen during weekends during holidays during u they have follow this pattern of human weaknesses no they don't hack computers these days they hack human emotions they hack human behavior so to break it I think an AI based system which can uh you know decide in split of second is would be a very helpful uh step simultaneously if it is backed by legal framework you know by enacting AI laws, AI governance rules or specific rules for doing this great work of cyber crime monitoring and cyber crime detection. I think that will stand uh its true test and it will hold uh in court for longer time.
24:49 >> That's right. Dr. Chakrabati, one question uh uh that that I would like to pose to you as an AI expert. uh do you also think there are some technological challenges uh in implementing say the AIEL fraud monitoring at scale which is required in our country across the banks and the digital payment systems. >> Yes. So one of the major challenge that I can see at this point of time is basically there should be human in the loop always >> because uh you cannot give the entire uh responsibility to AI >> uh because what happens uh this is a pattern recognition software and it will identify uh and probably put uh thousands and millions of alerts per day. Now for a analyst to actually I
25:35 mean uh to see which are the alerts which actually requires some kind of a attention for that I think that human in the loop is very very important. So uh and that is also at a scale because we we are talking about the entire financial ecosystem here. Second thing u what I can see is uh this AI need to be upgraded uh very frequently. Okay. So if you are using any kind of models or any kind of uh trained LLMs or SLMs or something like that. So that is also going to be a challenge because this the fraud stars will be coming up with newer and newer techniques. >> Absolutely. So we need to evolve with the change the challenges and since you spoke about human one last question Mr. Kumar uh we have very little time left but what role will the citizens play in
26:21 strengthening this ecosystem? Uh how important is timely reporting uh through platforms like the national cyber crime reporting portal? >> Very true. Just one short comment I would like to make on what Mr. Secretary has just said that I just want to assure that human will always be there in the loop. This system is just to automate the process and assist the human >> enabler. Yes, there is an enabler ultimately the decision will be taken by human only. That is one part and second you very rightly said that we are taking these preventive measures but of course a big responsibility lies on the citizens like what our panelist Dr. Prashant Mali has said that they are not hacking the system mostly in case of uh at least in case of cyber crime they are
27:09 you know hacking our human emotions or they are mostly you will find that technology has been exploited but it all has been started with social engineering. So we have to uh the first thing is that we need to aware and the citizens that what type of frauds are happening and everybody should know and remember that we have a national cyber crime helpline number which is 1 1930. So any type of cyber fraud they can immediately report on 1930 and second option is that they can always come on a portal and they can uh complain on www.pyerscrime.gov.in. So what happens as soon as the campaign is registered immediately the bank and the police they are all onboarded on the back end of these portals they immediately start taking action. So uh
27:57 the the early the citizens report the early action will be taken and then high probability that their money will be saved. >> That's right. Right. I think as India's digital economy grows uh the importance of securing it also becomes all the more important and and uh initiatives like these what we've discussed through the course of the program uh do clearly signal uh that that there is a decisive shift that India is making uh towards making the ecosystem uh more faster, more smarter and more secure. Time allows us to take up only that much on the program. Gentlemen, thank you so much for uh joining us, sharing your insights with us and our viewers. to you viewers. Thank you very much for your time as well.
28:53 To watch more programs of Sunset TV, subscribe to our YouTube channel and don't forget to like and share them.
