Deep learning is an advanced and flexible technology widely used in fields such as natural language processing, machine learning, and computer vision. One of its most recent applications is the creation of deepfakes, essentially manipulating videos or photos. While this technology holds a lot of promise, it is increasingly being used for malicious purposes, such as spreading fake news, creating misleading videos, committing financial fraud, and revenge porn. This makes public figures, including celebrities and politicians, particularly vulnerable to deepfakes. This article examines the basic techniques used to design and analyze deepfakes, focusing on various deep learning methods. It also highlights the limitations of existing methods and available research data. Since deepfakes can be easily created and distributed, the lack of robust detection tools poses a significant challenge. However, deep learning-based problem solving has been shown to be superior to traditional methods in solving this problem.