Synthetic media is digital content in various media formats, including text, image, and video, which has been automatically and artificially produced or manipulated. Although not all synthetic media is AI-generated, it often refers to the use of generative AI to produce content, such as deepfakes, through the use of artificial intelligence within a set of human-prompted parameters. Synthetic media as a field has grown rapidly since the creation of generative adversarial networks, primarily through the rise of deepfakes as well as music synthesis, text generation, human image synthesis, speech synthesis, and more. Though experts use the term "synthetic media," individual methods such as deepfakes and text synthesis are sometimes not referred to as such by the media but instead by their respective terminology (and often use "deepfakes" as a euphemism, e.g. "deepfakes for text" for natural-language generation; "deepfakes for voices" for neural voice cloning, etc.) Significant attention arose towards the field of synthetic media starting in 2017 when Motherboard reported on the emergence of AI altered pornographic videos to insert the faces of famous actresses. Potential hazards of synthetic media include the spread of misinformation, further loss of trust in institutions such as media and government, the mass automation of creative and journalistic jobs and a retreat into AI-generated fantasy worlds. Synthetic media is an applied form of artificial imagination.
History
Pre-1950s
The idea of automated art dates back to the automata of ancient Greek civilization. Nearly 2000 years ago, the engineer Hero of Alexandria described statues that could move and mechanical theatrical devices. Over the centuries, mechanical artworks drew crowds throughout Europe, China, India, and so on. Other automated novelties such as Johann Philipp Kirnberger's "Musikalisches Würfelspiel" (Musical Dice Game) 1757 also amused audiences. Despite the technical capabilities of these machines, however, none were capable of generating original content and were entirely dependent upon their mechanical designs.
Rise of artificial intelligence
The field of AI research was born at a workshop at Dartmouth College in 1956, begetting the rise of digital computing used as a medium of art as well as the rise of generative art. Initial experiments in AI-generated art included the Illiac Suite, a 1957 composition for string quartet which is generally agreed to be the first score composed by an electronic computer. Lejaren Hiller, in collaboration with Leonard Issacson, programmed the ILLIAC I computer at the University of Illinois at Urbana–Champaign (where both composers were professors) to generate compositional material for his String Quartet No. 4. In 1960, Russian researcher R.Kh.Zaripov published worldwide first paper on algorithmic music composing using the "Ural-1" computer. In 1965, inventor Ray Kurzweil premiered a piano piece created by a computer that was capable of pattern recognition in various compositions. The computer was then able to analyze and use these patterns to create novel melodies. The computer was debuted on Steve Allen's I've Got a Secret program, and stumped the hosts until film star Harry Morgan guessed Ray's secret. Before 1989, artificial neural networks have been used to model certain aspects of creativity. Peter Todd (1989) first trained a neural network to reproduce musical melodies from a training set of musical pieces. Then he used a change algorithm to modify the network's input parameters. The network was able to randomly generate new music in a highly uncontrolled manner. In 2014, Ian Goodfellow and his colleagues developed a new class of machine learning systems: generative adversarial networks (GAN). Two neural networks contest with each other in a game (in the sense of game theory, often but not always in the form of a zero-sum game). Given a training set, this technique learns to generate new data with the same statistics as the training set. For example, a GAN trained on photographs can generate new photographs that look at least superficially authentic to human observers, having many realistic characteristics. Though originally proposed as a form of generative model for unsupervised learning, GANs have also proven useful for semi-supervised learning, fully supervised learning, and reinforcement learning. In a 2016 seminar, Yann LeCun described GANs as "the coolest idea in machine learning in the last twenty years". In 2017, Google unveiled transformers, a new type of neural network architecture specialized for language modeling that enabled for rapid advancements in natural language processing. Transformers proved capable of high levels of generalization, allowing networks such as GPT-3 and Jukebox from OpenAI to synthesize text and music respectively at a level approaching humanlike ability. There have been some attempts to use GPT-3 and GPT-2 for screenplay writing, resulting in both dramatic (the Italian short film Frammenti di Anime Meccaniche, written by GPT-2) and comedic narratives (the short film Solicitors by YouTube Creator Calamity AI written by GPT-3).
Branches of synthetic media
Deepfakes
… excerpt ends here. Continue reading the full article.
