Synthetic Media: Research
Find an example of synthetic media (marketing, art, entertainment, ???)
AI Influencers
I think that the creation of AI influencers can be considered use of synthetic media because it is essentially taking CGI characters and inserting them into the social media landscape for the purpose of marketing. Although it is not really depicting the use of machine learning, I think the aspect of data training comes from the images and content that they create with the character using what marketing tactics draw the attention of users on social media. There are definitely apparent ethical issues with this type of digital ‘product’ because although they are not being used at that capacity yet, it mimics the social behavior of humans too much that it could soon pose danger on society and social media users.
Google Deep Dream
Another example of synthetic media can be digital artworks created through Google’s Deep Dream program. Through this program users can reprocess, arbitrarily manipulate and create entirely new images using machine learning capabilities. This program would be considered synthetic media as it takes information from the existing image and replaces, or adds onto the image elements from other images in a newly generated style and medium. The results that it creates can be quite hallucenagetic and obviously digitally produced which seems to be the objective of the program. I think that for this particular program, the ethical ramifications do not seem like a huge issue, because the output usually results in a considerable amount of manipulation even to very recognizable images to the point where it cannot really be compared.
GauGan
Neural Doodle
Programs such as GauGan and Neural Doodle allow the user to create simple sketches and drawing tools that then take information from data assets that have been programmed already. Users can also add new data to the asset to create different directions of output as well. Tools like GauGan and Neural Doodle demonstrate the use of machine learning in a really interesting and prolific way that allows the user to achieve a lot of output with little input. The user can manipulate the program to achieve a certain style of output they desire and combine the physical act of drawing and usage of machine learning. I think that the ethical implications for programs like these are trickier than other especially because some of the styles that they finished output mimics is those of significant artists that have solidified their body of work through their unique styles such as Monet or Picaaso. It brings to question how much value do these works created through programs like GauGan and Neural Doodle can hold coupled with the tricky boundaries of copying another artists style through a simple click or two. I also think that the reason these questions arise is because the output of these programs are such high quality and makes me wonder in what ways will they be applied by the users.
Cindy Sherman
Although not exactly an example of the use of machine learning, I think that the works of Cindy Sherman were early predecessors to the later created synthetic media arts. I say this because the process of creating synthetic media and art through machine learning heavily relies on creating drastically different outputs with the use of a single program or limited assets. In a sense her work embodies this idea but instead of having the aid of artificial intelligence, she reverts to analog methods such as prosthetics, special effects makeup, setting a stage, various methods of dramatic photography, and also photoshop in her most recent works. I think that the ethical ramifications of her work could lie on the subject matters she portrays and whether it borders on appropriation or not, but since she uses herself as the medium and transforms herself into the subject and it does not refer specifically to an existing individual, I believe that the ethical issues are minimal compared to digital artists who often times blatantly appropriate other images, cultures, or individuals.
︎︎︎ BACK TO ITP BLOG