Synthetic media
Synthetic media (also known as AI-generated media,[1][2] media produced by generative AI,[3] personalized media, personalized content,[4] and colloquially as deepfakes[5]) is a catch-all term for the artificial production, manipulation, and modification of data and media by automated means, especially through the use of artificial intelligence algorithms, such as for the purpose of misleading people or changing an original meaning.[6][7][8] Synthetic media refers to any form of media, including but not limited to images, videos, audio recordings, and text, that are generated or manipulated using artificial intelligence (AI) techniques. This technology enables the creation of highly realistic content that may be indistinguishable from authentic media produced by humans. 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.[8] 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.)[9][10] 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.[11][12] Potential hazards of synthetic media include the spread of misinformation, further loss of trust in institutions such as media and government,[11] the mass automation of creative and journalistic jobs and a retreat into AI-generated fantasy worlds.[13] Synthetic media is an applied form of artificial imagination.[11]
Not to be confused with a synthetic growth medium.Concerns and controversies[edit]
Deepfakes have been used to misrepresent well-known politicians in videos. In separate videos, the face of the Argentine President Mauricio Macri has been replaced by the face of Adolf Hitler, and Angela Merkel's face has been replaced with Donald Trump's.[87][88]
In June 2019, a downloadable Windows and Linux application called DeepNude was released which used neural networks, specifically generative adversarial networks, to remove clothing from images of women. The app had both a paid and unpaid version, the paid version costing $50.[89][90] On June 27 the creators removed the application and refunded consumers.[91]
The US Congress held a senate meeting discussing the widespread impacts of synthetic media, including deepfakes, describing it as having the "potential to be used to undermine national security, erode
public trust in our democracy and other nefarious reasons."[92]
In 2019, voice cloning technology was used to successfully impersonate a chief executive's voice and demand a fraudulent transfer of €220,000.[93] The case raised concerns about the lack of encryption methods over telephones as well as the unconditional trust often given to voice and to media in general.[94] AI-powered voice synthesis techniques allow for the creation of realistic synthetic voices, which can be used for dubbing, voiceovers, and personalized virtual assistants.
Starting in November 2019, multiple social media networks began banning synthetic media used for purposes of manipulation in the lead-up to the 2020 United States presidential election.[95]
Potential uses and impacts[edit]
Synthetic media techniques involve generating, manipulating, and altering data to emulate creative processes on a much faster and more accurate scale.[96] As a result, the potential uses are as wide as human creativity itself, ranging from revolutionizing the entertainment industry to accelerating the research and production of academia. The initial application has been to synchronize lip-movements to increase the engagement of normal dubbing[97] that is growing fast with the rise of OTTs.[98] News organizations have explored ways to use video synthesis and other synthetic media technologies to become more efficient and engaging.[99][100] Potential future hazards include the use of a combination of different subfields to generate fake news,[101] natural-language bot swarms generating trends and memes, false evidence being generated, and potentially addiction to personalized content and a retreat into AI-generated fantasy worlds within virtual reality.[13]
Advanced text-generating bots could potentially be used to manipulate social media platforms through tactics such as astroturfing.[102][103]
Deep reinforcement learning-based natural-language generators could potentially be used to create advanced chatbots that could imitate natural human speech.[104]
One use case for natural-language generation is to generate or assist with writing novels and short stories,[105] while other potential developments are that of stylistic editors to emulate professional writers.[106]
Image synthesis tools may be able to streamline or even completely automate the creation of certain aspects of visual illustrations, such as animated cartoons, comic books, and political cartoons.[107] Because the automation process takes away the need for teams of designers, artists, and others involved in the making of entertainment, costs could plunge to virtually nothing and allow for the creation of "bedroom multimedia franchises" where singular people can generate results indistinguishable from the highest budget productions for little more than the cost of running their computer.[108] Character and scene creation tools will no longer be based on premade assets, thematic limitations, or personal skill but instead based on tweaking certain parameters and giving enough input.[109]
A combination of speech synthesis and deepfakes has been used to automatically redub an actor's speech into multiple languages without the need for reshoots or language classes.[108] It can also be used by companies for employee onboarding, eLearning, explainer and how-to videos.[110]
An increase in cyberattacks has also been feared due to methods of phishing, catfishing, and social hacking being more easily automated by new technological methods.[94]
Natural-language generation bots mixed with image synthesis networks may theoretically be used to clog search results, filling search engines with trillions of otherwise useless but legitimate-seeming blogs, websites, and marketing spam.[111]
There has been speculation about deepfakes being used for creating digital actors for future films. Digitally constructed/altered humans have already been used in films before, and deepfakes could contribute new developments in the near future.[112] Amateur deepfake technology has already been used to insert faces into existing films, such as the insertion of Harrison Ford's young face onto Han Solo's face in Solo: A Star Wars Story,[113] and techniques similar to those used by deepfakes were used for the acting of Princess Leia in Rogue One.[114]
GANs can be used to create photos of imaginary fashion models, with no need to hire a model, photographer, makeup artist, or pay for a studio and transportation.[115] GANs can be used to create fashion advertising campaigns including more diverse groups of models, which may increase intent to buy among people resembling the models[116] or family members.[117] GANs can also be used to create portraits, landscapes and album covers. The ability for GANs to generate photorealistic human bodies presents a challenge to industries such as fashion modeling, which may be at heightened risk of being automated.[118][119]
In 2019, Dadabots unveiled an AI-generated stream of death metal which remains ongoing with no pauses.[120]
Despite the risks, synthetic media also presents innovative opportunities for artistic expression, storytelling, and entertainment.
Musical artists and their respective brands may also conceivably be generated from scratch, including AI-generated music, videos, interviews, and promotional material. Conversely, existing music can be completely altered at will, such as changing lyrics, singers, instrumentation, and composition.[121] In 2018, using a process by WaveNet for timbre musical transfer, researchers were able to shift entire genres from one to another.[122] Through the use of artificial intelligence, old bands and artists may be "revived" to release new material without pause, which may even include "live" concerts and promotional images.
Neural network-powered photo manipulation has the potential to abet the behaviors of totalitarian and absolutist regimes.[123] A sufficiently paranoid totalitarian government or community may engage in a total wipe-out of history using all manner of synthetic technologies, fabricating history and personalities as well as any evidence of their existence at all times. Even in otherwise rational and democratic societies, certain social and political groups may use synthetic to craft cultural, political, and scientific cocoons that greatly reduce or even altogether destroy the ability of the public to agree on basic objective facts. Conversely, the existence of synthetic media will be used to discredit factual news sources and scientific facts as "potentially fabricated."[55]