In 2025, Kaspersky Premium for Windows received the annual ‘Approved’ certification by leading testing lab AV-Comparatives. This was given to Kaspersky Premium for detecting 93% of phishing URLs, demonstrating strong anti-phishing capabilities powered by AI technology and a commitment towards safeguarding users’ digital life and privacy.
Phishing remains one of the most dynamic online fraud threats, constantly evolving and becoming more complex in light of new technology. The active use of artificial intelligence (AI) by cybercriminals allows them to create fake email and web pages faster and generate plausible audio and video content, thus making phishing attacks targeted and even more sophisticated. This necessitates increased caution on the part of users, as well as tools which are highly effective at the detection of phishing attempts, such as Kaspersky cybersecurity solutions which have consistently demonstrated advanced capabilities in this field.
In the 2025 evaluation by AV-Comparatives, a range of security solutions were tested with more than 200 phishing URLs. To complete certification, a product needed to block at least 85% of these phishing sites. Additionally, a false-positive test was performed to verify that legitimate websites, particularly sensitive ones such as online banking, were not misguidedly blocked.
Since 2011, Kaspersky’s consumer solutions have regularly participated in Anti-phishing tests by AV-Comparatives and have demonstrated stable, high results. In particular, Kaspersky achieved a 93% detection rate with 0 false positives in 2025 and 2024, and was awarded with the “Approved” certificate.
Anti-phishing protection powered by AI
To ensure advanced cyber protection against increasingly complex phishing threats, Kaspersky actively amplifies its consumer solutions with AI-powered tools. Kaspersky’s anti-phishing technology has two machine learning-based algorithms, which make phishing detection effective and easily adaptable to new types of phishing. In 2024 more than 67 million phishing attacks were blocked by Kaspersky Consumer solutions with the help of AI technology.
In particular, Kaspersky’s anti-phishing protection is run by the AI-powered feature OCR (Optical character recognition) which is indispensable in recognising malicious text hidden inside pictures on a phishing site. This machine algorithm analyses all downloaded web pages and uses machine vision to detect phishing text inside the image. Every discovered phishing text is saved in the system and used for subsequent detection.
Another technology that makes the Kaspersky Anti-phishing solution highly effective is its patented machine learning model based on the examples of multiple phishing and legitimate sites. The algorithm is trained on millions of downloaded web pages, extracting from them those trappings that are the most relevant for detecting fraudulent web pages.
“Leveraging machine learning in anti-phishing tools transforms the phishing detection approach, by enabling our solutions to detect the threats that are too complex for more traditional approaches, empowering our customers to stay one step ahead. The high results in AV-Comparatives anti-phishing annual certification showcases the advanced protection our security solutions provide, giving users an extra layer of digital security,” comments Vladislav Tushkanov, Group Manager at Kaspersky AI Technology Research Center.
To be highly protected against phishing attacks Kaspersky recommends:
- Always check the authenticity of unsolicited emails and links, even if they seem trustworthy.
- Avoid sharing personal data on unverified sources.
- Double-check websites before making purchases, especially when the offer seems too profitable.
- Use a reliable security solution to block phishing attempts.
