Yandex Browser Detects Over 400,000 Phishing Sites Using AI

In the past five months, Yandex Browser has identified more than 400,000 fraudulent websites with the help of artificial intelligence (AI). This data comes from company statistics reviewed by Vedomosti. According to Yandex Browser, Russian internet users attempt to access phishing sites about 1.5 million times every day. In 90% of cases, these attempts are made from mobile phones, including through messengers, emails, and apps.

Yandex has been testing AI-based phishing site detection since November 2023. The AI checks websites before allowing users to access them. The algorithms analyze factors such as the website’s creation date, the amount of time users spend on the site, visit frequency, and whether the site appears in search results or is accessed via another service.

“For example, if a company suddenly has a new one-page website that was created just a few hours ago and is registered to a private individual instead of the company, the algorithms may flag it as phishing,” explained a Yandex Browser representative.

Yandex claims its phishing site detection accuracy is 99.9%. The company reports that the algorithms find over 1,500 such sites daily, but only one or two turn out to be genuine phishing threats.

Currently, the neural networks are learning to identify new types of phishing. Recently, phishing sites have increasingly been imitating banking apps, cryptocurrency exchange platforms, and dating services that have been removed from app stores. “Fraudsters are very inventive. We have even found scam pages offering help to victims of online fraud,” noted the Yandex Browser team.

Earlier today, it was reported that the number of legal disputes involving AI has increased by 60%.