AI deepfake technologies

The development of artificial intelligence has led to major technological advancements but has simultaneously created new ethical and security challenges, most notably the emergence of deepfake technology stands out. Based on neural networks, this technology enables the creation of highly realistic yet fabricated audio-visual content, threatening information integrity and public trust. This paper presents a systematic analysis of recent scientific and professional literature, focusing on deepfake technologies and their ethical implications. Sources include databases such as PubMed, Google Scholar, IEEE Xplore, and leading journals in information technology and ethics. Researches from 2025 indicate a drastic rise in both the volume and sophistication of deepfake content, posing growing risks to online security. Deepfake files are projected to reach 8 million by the end of 2025, compared to half a million in 2023. Furthermore, human detection accuracy for high-quality deepfakes remains low, averaging only 24.5%. This study aims to examine models, algorithms, and strategies for their identification and mitigation.