New Delhi. In the Internet world, it is no longer easy to identify whether the user of a website is a real human being or whether an Artificial Intelligence (AI) program is working in his place. Earlier, the CAPTCHA checkbox provided on websites was considered sufficient to confirm that the user is not a machine. But now modern AI technology has challenged even this security system. Due to this, Indian banking and e-commerce companies are paying attention to new cyber security measures. Especially the automatic transactions done through AI agents have increased the concern of the companies involved in digital business.
Why is the old security of CAPTCHA failing?
The main purpose of CAPTCHA was to protect websites from automatic bots. In this, tasks like identifying pictures, reading letters or clicking on checkboxes are given. But new vision-language and multimodal AI models are now able to solve such tests faster. According to information revealed by Akamai Technologies, AI systems can solve many types of visual and text-based CAPTCHA tests as fast as or even faster than humans. Due to this, the reliability of old security methods is decreasing.
Giving banking access to AI agents can be a danger
Another serious aspect of this challenge is that some people are giving their AI agents access to banking accounts and digital services. When an AI bot gets permission to use OTP or account login information, it can be difficult to establish responsibility for the transactions done through it. Now it has also become important for websites to understand whether the traffic coming to their platforms is from real customers, from an authorized AI agent or is part of the activity of a cyber criminal.
Now identification will be done based on browser and typing method
Technology companies are now developing multi-layered security systems instead of relying only on CAPTCHA. Under this, browser behavior, typing method, mobile biometrics and activities happening in the background will be analyzed. This can help identify suspicious automated activities. The purpose of such systems is to block suspicious bots without disturbing real customers.
Zerodha also expressed concern
According to Kailash Nath, CTO of Zerodha, it is technically extremely difficult to identify AI agents or stop them completely. He says that when the user himself gives complete access to the OTP or account to his AI agent, it becomes complicated to identify the party responsible for any wrong transaction later.
MasterCard and Visa are introducing new security measures
To find a solution to this problem, MasterCard is working on Agent Pay Framework in India. Through this system, transactions done by certified AI agents can be traced with the help of cryptographic tokens. Similarly, Visa is also working on its Trusted Agent Protocol. The purpose of these systems is to ensure that in every digital transaction it is clear which person or AI agent has carried out it.
Increased difficulties for e-commerce companies and small businessmen
Due to the increasing activity of AI bots, e-commerce companies may have to face many problems. Increase in unnecessary traffic on the website puts pressure on the servers and may delay the delivery of services to customers. Apart from this, fake orders of cash on delivery and unnecessary returns can increase the expenses of the traders. Small businesses also do not have enough resources to install expensive cyber security systems. Companies like Zomato and Blinkit are also keeping an eye on this changing threat. With the increasing use of AI agents in the future, customer identity and transaction security may become the biggest priority for digital businesses.