Online social media is the dominant force in today's world in many ways, and its user base is expanding daily. Social media usage is rapidly increasing. The main advantage is the ease and effectiveness with which we may communicate with others through online social media. This created a new possible assault vector, including a fake identity, misleading information, and so forth. Based on a recent survey, there are far more accounts on social media than users within the population. Online social network providers find it challenging to identify these fraudulent accounts. Recognizing these fraudulent accounts is crucial because social media is overrun by advertisements, misleading information, and other kinds of content. In this research, we present a strategy that uses advanced machine learning strategies, like high gradient boosting algorithms, to increase the accuracy of recognizing false profiles. Our approach represents an important step toward a safer and more reliable cyber world, where the age-old problem of fictitious social media profiles is being addressed in a very practical way. The fact that so many people use social media today gives one an idea of the immediacy with which it is necessary to take some serious action against the growing number of fake profiles. Such profiles compromise not only the reliability of online communities but also pose serious security threats to users’ private information. Our idea will help in building a safer virtual world where people can connect on social media platforms with confidence and without fear of being deceived or exploited.