Safe URL Detection with Privacy Using Machine Learning and Cryptography Techniques

A malicious URL or website represents a serious cybersecurity threat. Search engines logically serve as the foundation for managing in order. However, the influx of many malicious websites on search engines has created a considerable threat to our users. Most existing systems designed to identify malicious websites concentrate on particular types of attacks. Simultaneously, available browser extensions relying on blacklists are ineffective against numerous websites. To stop the server from deriving any useful information from the masked data, any data supplied from the client side must be successfully hidden. We provide the first Privacy-Preserving Safe Browsing service in this regard. This service offers strong security guarantees now absent from other Safe Browsing services. In particular, it preserves the capacity to recognize dangerous URLs while safeguarding the user's privacy and the blacklist provider's proprietary assets. In this study, we provide a methodology that protects users' privacy from outside analysts and service providers by encrypting their perceptual data. Furthermore, while maintaining differential privacy, it fully enables selected aggregate operations for examining user behavior online. The AES (Advanced Encryption Standard) algorithm encrypts users' online search behavior data. Only authenticated users can decrypt their history with assistance from the secret key sharing and verification process. A user feedback system notifies the search engine administrator of any malicious URLs, enabling the administrator to update the malicious URL database frequently.