Image source: Getty Images

The situation is like that scene in a cowboy movie, in which two people are standing face to face and the question is who will take out the gun first.

On one side there are hackers and on the other side there are IT teams of banks.

The battle between the two is about what kind of Artificial Intelligence (AI) the hackers will use for cyber theft in the banks and how quickly and effectively the banks will be able to use their AI tools to stop them.

According to cyber security experts, this is the biggest challenge facing the banking system today.

The biggest reason for this is the 'machine-like speed' provided by AI.

Cyber ​​security experts have warned that a new flaw revealed in the IT system may affect the old technical infrastructure. Especially India's banking system can be affected by this.

Experts say that cyber attacks may start by next month, but even before that.

Cyber ​​attack very quietly

Image source: AFP via Getty Images

Rahul Sasi, co-founder and CEO of CloudSec, wrote on social media, "This month is going to be even more hectic as a new flaw in Java called Java to Shelch has been revealed."

End

Special stories selected for you by Nitin Srivastava

Learn more

"This can affect the most legacy technology systems, such as the banking system. Most of the systems of banks run on Java."

Rahul Sasi told BBC News Hindi, "Many new methods of attack are emerging. The flaws which humans or machines could not find earlier, are now being discovered by AI."

"These vulnerabilities allow access to any computer or server. Any type of system you can think of can be accessed."

But it is not only Rahul who is giving such warning.

Rajshekhar Pullabatla, founder of BreachX, told BBC News Hindi, "Rahul Sasi is right. Many times, incidents of service outages are dismissed as normal technical glitches."

"In fact, many of these incidents are caused by cyber attacks in which the attackers were not able to enter the system, but they did shut down the system."

"This is only the part that we see. The cyber attacks that hackers are successful in happen very quietly."

Shashidhar CN, Founder, Security Consultancy Services LLP also agrees with these two experts.

However, the spokesperson of Indian Banks Association i.e. IBA refused to comment on this issue.

He told BBC News Hindi, "Will not be able to comment right now."

Is it a technical flaw or a failed cyber attack?

The services about which Rajashekhar talked about closure were also mentioned in Rahul Sasi's social media post.

Rahul Sasi wrote, "Warning: The incidents of closure of services that are being seen in the world are not happening suddenly or without any reason. All these are related to cyber attacks and they show how broken the internet system is."

He says that it is not always necessary for hackers to be successful in entering a system to cause damage.

Sometimes even a failed attack can crash an app or shut down the entire service.

Rahul Sasi gave the example of PHP2Shell i.e. CVE-2026-109618 for this.

He wrote, "Many of the publicly available attack methods using this flaw were not reliable enough to remotely execute arbitrary code on the server."

"But they could crash the server."

Sassi said that as efforts to exploit these vulnerabilities increase, an incident may appear to be a simple technical glitch when it may actually be a failed cyberattack.

He says, "The attacker failed to enter the system, but the system still shut down."

When Rahul Sasi was asked why the situation was going to get more difficult, he said, "Now more and more flaws will come to light, which have never been discovered before. All these things will come to light simultaneously."

Why has AI become so dangerous?

Shashidhar says that earlier it used to take 24 hours to fix any serious defect.

If the threat was of medium level, one month's time could be given to fix it.

The low-risk flaw was said to be fixed in the next software update.

But now the situation has changed.

Shashidhar says, "Now 24 hours is a very long time even for a medium level threat. During this time our system may already be vulnerable."

"Someone could take advantage of that flaw with the help of AI."

Rajashekhar Pullabatla says that for the last 20 years the emphasis of cyber security has been on detecting attacks.

In this method, the attack is first waited for, then it is identified and then a response is given.

But this method works only when that type of threat has been seen before.

Rajasekhar says that attackers using AI can discover flaws that have never been seen before.

In such a situation, the security system does not have any prior information to identify it.

He says, "The field of cyber security now has to move beyond detecting attacks to finding vulnerabilities. That means finding vulnerabilities before the attacker can."

The biggest challenge to stop the attack is speed

Image source, Ishant Chauhan/Hindustan Times via Getty Images

According to Shashidhar, the real competition is how fast banks can protect their systems and how fast hackers can break into them.

Earlier the time difference between the two was more. AI has reduced this gap a lot, because now many tasks can be done automatically and very quickly.

But it is not easy to fix the flaws in the banking system on your own.

Shashidhar gives the example of the messages that banks send to their customers.

In these messages it is told that banking services will remain closed for half an hour or an hour late at night for 'maintenance'.

He says, "If any flaw has to be fixed, then the entire system has to be stopped. The app has to be closed and then the flaw has to be fixed."

Due to this process the speed of repairing the system decreases.

Shashidhar says, "These are ground challenges. There are no rules for hackers. One small mistake on your part is enough for them."

"But even a small thing can become a big problem for the person protecting the system. This is the real challenge. AI has become a kind of monster."

Rajashekhar goes one step further than this.

He says, "With the help of AI, we have seen this work done in about an hour."

"Attacks now happen at the speed of a machine. Therefore, as soon as any flaw is detected, preparations should be made to avoid it."

According to Rajasekhar, many banks are still using security technology created 20 years ago.

Hackers who attack with the help of AI look for such old systems.

Where might the next attack take place

Rajasekhar says, "An attack detection system tells you that you have been attacked. A vulnerability detection system tells you where you might be attacked."

"Only the second of these methods gives you time to act."

He says that the security team of a company should not wait for a warning from the software company to know from where its system can be breached.

How long does it take to fix a defect?

Rajashekhar explains about the 'zero-day' flaw by giving the example of his company.

Zero-day is a flaw that even the company making the software is not aware of. Therefore, there is no available way to fix it.

An attack using this vulnerability can easily bypass security measures that are designed to prevent only known threats.

According to Rajasekhar, BreachX Zero-Day Labs has discovered more than 1,700 zero-day vulnerabilities in widely used software.

Information about only some of these deficiencies has been posted on the company's website.

The reason for this is that if information about a flaw is made public before it is fixed, attackers can take advantage of it.

Rajasekhar says, "Now the speed of finding flaws has become greater than the speed of fixing them."

"We have over 1,700 zero-day vulnerabilities. Only 11 of these have been fixed in the last four months."

"AI is finding bugs at the speed of machines, while the world is still fixing them at the speed of humans."

'Virtual patch' until real fix arrives

It may take several weeks for a software company to fix a flaw.

But the bank cannot shut down its systems for so long.

Rajasekhar's company uses a 'virtual patch' in such situations.

This prevents attacks from occurring through that vulnerability until an actual fix is ​​implemented.

Is AI the real villain?

Rajashekhar does not agree with this.

He says, "AI is not a villain. With the same power with which it can discover flaws, it can also fix them."

"But a team of humans alone cannot keep up with the speed of AI-assisted attacks. Defense will also have to be done at the speed of machines."

Rajasekhar once again points out that many banks are still using security technology created 20 years ago.

According to him, hackers who attack with the help of AI look for such old systems.

What should banks do to protect customers?

Rahul Sasi says, "This is the right time to strengthen the security and technical infrastructure."

According to him, banks will have to fix the flaws faster than before and increase their security.

He says, "This is the right time to further strengthen security."

What should bank account holders do?

"Don't rely on just one bank account and one debit card," says Rajasekhar.

"If your account is frozen during an investigation, it may remain down for a period of time."

He advised to open an account in another bank and keep enough money to meet expenses for one or two months.

According to Rajasekhar, the holiday season at the end of the year is more dangerous. Therefore this is the right time to do so.

Published by Collective Newsroom for the BBC.

Download the new BBC World Service app

Choose BBC News Hindi and get news alerts, live TV, podcasts... all in one place

Please note: The new app is not available in the UK and US.

short video

end of short video

top story

Has a cyber war started between AI-equipped hackers and IT teams of banks?

Live, on Trump calling Putin a 'friend', Zelensky said - 'He is a murderer'

Why did Congress oppose the visit of Pakistan Minister Musadi Malik to India?

must read

most popular