Nowadays, with the continuous development of technology, Digital banking, e-tech and social commerce are all booming their existence, bringing along all sorts of actions of financial frauds, which also thrive continuously through the pass of times.

On September 20, a new anti-fraud report seminar was held in Beijing. During the event, Wu Zhong, general manager of DataVisor China, and Yang Hao, senior industry consultant, introduced the latest anti-fraud report in 2019, which focused on the highly concerned account theft issues, and introduced relevant risk control problems and technical solutions to more than 10 media agencies at present.

Internet thefts: not to be underestimated

At this age, fraudsters are not limited to making profits with a stolen credit card, but instead, they are more into forming a complete ecosystem, including credit card theft, personal information capturing, privacy authentication and others as a series of processes.

“After stealing, it can not only transfer, remit and trade directly, but can also sell the stolen information for a second time and make profit illegally.” Wu Zhong believes that the current problem of account theft is becoming more and more serious, resulting in huge economic losses. The so-called account theft (ATO) refers to the attacker stealing the account password, login information to obtain the account, thus to retrieve personal information.

Take “account stealing” as an example. For fraudsters, account theft is profitable because it contains all kinds of valuable information, which can be used to send spam or commit other frauds. Moreover, with the explosive growth of data leakage cases, it becomes easier and cheaper to buy information of stolen accounts.

AI Unsupervised learning: effectively detects unknown risks

With the spreading application of 5G, the development of mobile Internet has reached its top, and the security problem has yet broken through the original boundary. The opportunities and challenges brought by artificial intelligence to the security industry have become one of the key issues that attract attentions.

It is reported that Datavisor, as one of the top 100 enterprises in the field of AI, implies automatic AI technology to combat against the intelligent network frauds. Wu Zhong introduced that Datavisor focuses on the security field as its core business, by using the original AI unsupervised machine learning algorithm to provide risk control solutions for Internet enterprises and financial institutions, reduce the risk of fraud attacks on enterprises, maintain the economic interests and reputation of enterprises, and contribute to the construction of a healthy and secure network environment.

The sparkling point of Datavisor is its original unsupervised machine learning technology, which is used to detect and intercept false traffic for enterprises, so as to help maintain the economic interests of enterprises, establish the reputation of enterprises, guarantee the timely and effective strategy of enterprises’ business process, and assist enterprises to obtain actual users. While facing such rampant acts of account theft cases and its threats, Yang Hao introduced some most popular risk control technologies in the world: black and white list, rule engine, supervised machine learning and unsupervised machine learning.

Yang Hao believes that compared with the first three, unsupervised learning has the advantage of effectively detecting some unknown risks or new types of fraud without labels. It is reported that Datavisor can find problems while fake users are conducting registration with an accuracy rate of 95%. Up to now, Datavisor has handled more than 800 billion user events globally, detected more than 200 million bad users, and protected more than 4.2 billion users from large global Internet companies.