Sentetik Medya
Visual, sound or video content that is partially or fully manufactured with artificial intelligence.
Synthetic media is a broad concept that covers visual, sound, video or text content partially or fully manufactured by AI algorithms. This category includes art works created with manufacturer traction networks (GAN) or diffusion models, such as human faces, visual-generating systems (such as Midjourney, DALL-E, Stable Diffusion), and a wide range of productions, such as sound clone technologies that mimic the voice of a real person, and video contents with completely artificial speakers. The main feature that distinguishes synthetic media is that content is drawn with a camera, which is statistically "produced" from the images that a model learned.
Synthetic media provides many creative and commercial benefits from the production of customized content in advertising, visual effects in the film industry, multilingual voice in educational materials. However, the same technology can also be abused to produce misleading content (deepfakes, fake news, imitation sounds), which is increasingly difficult to distinguish from the fact; due to this risk, content authentication standards (watering and resource labeling attempts such as C2PA) and synthetic content of platforms are labeling obligations, one of the most active discussion areas of today’s AI policy.
Synthetic media provides many creative and commercial benefits from the production of customized content in advertising, visual effects in the film industry, multilingual voice in educational materials. However, the same technology can also be abused to produce misleading content (deepfakes, fake news, imitation sounds), which is increasingly difficult to distinguish from the fact; due to this risk, content authentication standards (watering and resource labeling attempts such as C2PA) and synthetic content of platforms are labeling obligations, one of the most active discussion areas of today’s AI policy.