The problem of dipfakes is becoming more and more relevant at the moment. Dipfakes are considered to be unreliable information created or modified by the use of artificial intelligence technologies. The problem of dipfakes creates additional difficulties for validation of information reliability, as dipfakes methods allow to hide traces of interference or artificial creation of some information. This paper makes a complex research of the stated problematics. Within the framework of the article a classification of different methods of falsification detection on video image was made. In addition, a comparative analysis of these methods is carried out to identify the features and possibilities of their use for solving real-world problems. As a result, the presented information is systematized and conclusions are drawn concerning the possibility of using the presented methods for solving certain problems.