AI and Art
Artificial intelligence has a somewhat defined role in many industries and applications; it makes a lot of our work more efficient, tasks easier, transportation faster, and communication wider. From autonomous vehicles, medical diagnosis, to algorithmic trading, artificial intelligence has helped firms and establishments reach many technological feats. While artificial intelligence has found its home in various, obviously, tech-heavy industries, it is now beginning to be applied more readily in the arts. Although these applications may initially seem to contradict the notions of art and artistic creativity, it is indeed aiding artists in accomplishing feats never seen before.
Artists are now applying artificial intelligence to many of their mundane tasks and using it as an aid in tedious tasks. For instance, in the film industry, animators are working with engineers to train deep learning algorithms that help in cleaning up visual effects and making animated characters look more realistic [1]. Darren Hendler, the Digital Effects Supervisor at Digital Domain – a visual effects company, believes that artificial intelligence is helping artists and animators focus on more “creative tasks” as it undergoes “manual, mundane jobs.” On one of Hendler’s latest works, he led a team that applied artificial intelligence techniques to create Thanos for Avengers: Infinity War.
The team trained their software on high-resolution facial scans of Josh Brolin, the actor playing Thanos, to replicate his expressions to the greatest detail – even “down to individual wrinkles.” After that, they used another algorithm to plot Brolin’s expressions onto Thanos, leaving final artistic touches to the animators. As opposed to traditional face mapping techniques that could easily take weeks to complete, the machine learning software developed by Digital Domain performs the same task in “nearly real-time.” Similarly, Scriptbook, a tech-startup in Belgium, developed an AI algorithm that can read through a movie’s script and predict how well the movie will perform in the box office. The startup boasts that its algorithm is three times more successful than human readers in predicting the commerciality of a film and has predicted 22 of Sony Pictures’ box office failures, only in the last three years.
Artificial intelligence algorithms are not only aiding artists in their works, but are also creating their own artistic work. AI startup Amper Music developed an artificial intelligence algorithm that composes original music available to anyone, both beginners and pros [2]. As an example, Amper helped YouTuber Taryn Southern release the very first album that is entirely composed by AI software [3]. The algorithm takes in a few variations such as genre and tempo to compose entirely original music that can also be altered according to the artist’s taste.
As a result of AI-produced works, many questions have been proposed on the value of AI-produced works such as: “when an AI-composed song wins the Grammy, who gets the trophy?” [4] With artificial intelligence, artistic creativity enters a gray area where people do not know, firstly, if the mechanized work is worthy of credit and appreciation, and secondly, who to credit and appreciate? Moreover, as AI software continues to create and help create art works, the concerns faced by introducing AI in the arts are mainly centered on artistic creativity, not an AI domination of the industry that results in artists losing their jobs. As demonstrated in the film industry, artists aren’t losing their jobs to AI, but instead AI is aiding artists in mundane tasks, and leaving them to complete more creative jobs.
Likewise, in the music industry, AI, at the moment, will not outperform human artistic creativity, as AI-produced music does not have any context to it – artists do not only provide songs, but they also provide stories, cultures, and viewpoints that AI cannot provide [5]. Even if AI can include narratives with their musical composition, record labels are going to need to convince buyers that an AI story is more compelling than a human’s story. Hence, AI is merely a tool that helps artists compose original music faster and gives them more chances to express their creativity. Ultimately, artificial intelligence is helping artists across different industries perfect their craft in a faster and more efficient way. While concerns still exist in regards to the value of work created by software, all we can do now is be fascinated with what we have been able to achieve technologically and work to aid humans in more industries.
[1] Dan Robitzski, “Artificial Intelligence is Automating Hollywood. Now Art Can Thrive”, on Futurism.
https://futurism.com/artificial-intelligence-automating-hollywood-art/
[2] Ampere Music.
[3] Keith Nelson, “Taryn Southern’s new album is produced entirely by AI”, on Digital Trends.
https://www.digitaltrends.com/music/artificial-intelligence-taryn-southern-album-interview/
[4] David Pogue, “Is Art Created by AI Really Art?” on Scientific American.
https://www.scientificamerican.com/article/is-art-created-by-ai-really-art/
[5] Cherie Hu, “How Music Generated by Artificial Intelligence is Reshaping – Not Destroying – The Industry”, on Billboard.
Users who have LIKED this post:
6 comments on “AI and Art”
Comments are closed.
Fascinating blog post! As someone who loves to try and figure out which effects are practical and which are digitally enhanced. The early days of CG were filled with human oversight, but through the use of these new AI algorithms, the line between what is real and what is fake becomes more and more blurred.
Hi Rakanb, I like your article. When I heard about art, especially in a museum, they almost always have human characteristic words such as “passion, love, anger, depression, etc.” These emotions help trigger the artist to develop something that we might not of been exposed to without.
Do you think AI can ever replicate these emotions that are the backbone of certain art forms?
I personally think the extent of AI in generating ‘art’ will be limited to things like special effects and textures. As you said, a key component of art is the expression of human emotions and experience, which is something that, as far as we know, AI isn’t capable of achieving. When it comes to other areas like music, AI can likely generate something that is technically correct in terms of key, chord progression, and rhythm, but will lack the human component of storytelling. A good analogy would be the current application of neural networks in generating artwork or poetry (a quick Google search should reveal what I’m talking about). Of course it’s up to individual preference to determine what is and isn’t “good art”, but my perspective is – it may be interesting from a technical standpoint, but it falls short when it comes to human connection.
Interesting article, thanks! I think you’ve raised a really good point here – as beneficial as AI will be for companies to automate their decision-making processes, a lot of technological success over the decades has stemmed from the entertainment industries, and it is actually their drive to better technology that has resulted in flow on effects to other industries. Gaming, for example, greatly advanced computer processing power using the GPU, such that the GPU is now used instead of the CPU in many analytics tasks. In addition to this, film and entertainment tend to be hot-beds of innovation, with a massive pool of creative (and often technical) talent working in this fields. The film and media industries have so much revenue that I suspect technological advances may continue to stem from here, as well.
Really fascinating post – certainly begs the fundamental question of how we define art. Moreover, who receives ‘credit’? Interestingly enough the latter has also been debated for centuries. Akin to Hendler believing artificial intelligence helps artists focus on the more ‘creative tasks’ rather than the ‘manual, mundane jobs’, the most renown artists in modern history often had scores of apprentices and workers doing vast parts of the masterpieces we attribute to one name. For example, Peter Paul Ruebans, a Flemish artist would often leave the landscape and bodies to his studio members, and then sweep in at the end to bestow his artistic genius on the faces. The former was considered a learned skill – which anyone could do if well taught, but the facial expressions was where his creative sensibilities were thought to be best put to use. Similarly, Rodin and Warhol had large studios and most of their pieces would also be technically deemed collaborative efforts. Yet, they receive credit on account of their innovation and groundbreaking style. So is the idea enough? Or is there merit in the actual doing of the thing? Without ever picking up a paintbrush, if I commission a painting and describe exactly what I would like the artist to paint, the colors I want, how I want to feel when I look at it, and the subject matter and composition, do I now become the artist? AI generated music will certainly blur creative boundaries (though one might argue that they have always been blurred), but I echo David Pogue’s question – “What happens in a world where effort and scarcity are no longer part of the definition of art?”
I’d recently visited an Artists and Robots exhibition at the Paris Grand Palais (https://www.grandpalais.fr/en/event/artists-robots), and its impact was particularly chilling for me — there were bots that had been programmed to repeatedly draw from pictures and the emotions conveyed were very different from what could be felt in other museums, say, the Louvre. At the end of the exhibition they also questioned how far robots could go to create art that could be considered their own, or if any produced works would always go to the owner / creator of the robot.