In 2024, Yandex used artificial intelligence to speed up the process of finding and blocking advertising from fraudsters by 12 times. Attackers often change the content of advertising to phishing after passing moderation, trying to steal user data using QR codes.
Fraudsters and unscrupulous advertisers try to circumvent the rules and replace the content of sites with phishing. To combat this, Yandex uses artificial intelligence. It constantly checks ads and makes sure that the topic of the site and ads has not changed.
To solve this problem, several dozen different ML models are used today, which are constantly being retrained to improve the quality of search and the speed of blocking fraudsters' accounts. By the end of 2024, it has increased 12 times.
In 2024, the Yandex Direct system, designed for placing contextual advertising, blocked more than 300 thousand advertiser accounts that violated the placement rules. This is 40% more than in the previous year.
Over the past year, Yandex has blocked more than 20 thousand websites, about 4 thousand applications and approximately 29 thousand Telegram channels due to violations of the platform rules.
One of the main reasons for blocking advertising was phishing — more than 3 thousand domains were blocked due to such violations. Attackers often tried to steal user data using QR codes, as reported by Yandex.
In 2024, Yandex's automatic number detection system processed about one and a half billion calls from unknown numbers. Of these, 800 million were recognized as "undesirable". Compared to 2023, the share of calls from fraudsters increased by 16% and amounted to 56% of the total number.
Read materials on the topic:
Spying and stealing data: a new malicious program Zagrebator has been discovered in Russia
