I’ve always been someone who considers art one of my hobbies, enjoying both painting and creating digital art on Adobe programs in my spare time. While some of my friends and family might consider me #artsy, I would beg to differ that my computer can be even #artsier.
We have learned, discussed, and tweeted some about artificial intelligence already this semester, especially how artificial intelligence is disrupting certain industries. And while artificial intelligence is programmed to think like humans and respond the way Alexa does, what I find interesting is how artificial intelligence can actually create works of art just as exquisite and captivating as art made by humans. I was inspired by this topic after watching the WSJ Tech video titled Art By Artificial Intelligence.
Art created by computers means more than just images, but also music and literature. But focusing first on visual art, computers currently require existing images in order to create new works of art. Google created Deep Dream, which takes inputted images or photographs and uses AI to recreate an image with certain objects that the computer believes should be included, such as eyes and animals. The more images inputted to the computer, the better the computer becomes at interpreting images. I personally think it’s a hard concept to comprehend and the images appear to end up looking a little “trippy” as if you are dreaming, hence the name ‘Deep Dream.’ But what amazes me is that a computer can even take an image and alter it in a way to create a new type of art. So much so that 29 different Google Deep Dream artworks were auctioned in a San Francisco art show last year, selling for up to $8,000, which was more than any of the other art auctioned. Therefore, not only are computers creating what some might call a new type of art, but they are also outselling and competing against human artists, potentially disrupting the future of the art industry. Below are some Deep Dream examples:
(Inspired by Van Gogh’s The Starry Night)
If you are interested in viewing more images, click here for Deep Dream’s public library.
Adobe is another company experimenting with AI to create art, but in a different way. Engineers have created an AI program called Wetbrush that acts like a paintbrush to create digital art look like realistic oil paintings. Artists paint on a device using a stylus, but the stylus takes into consideration brush strokes, angles, and even time to dry; therefore, the stylus feels and acts like a real paint brush. The benefits I see in this type of AI is the ability to create masterpieces without the worry of smudging or messing up on paper, along with preventing the waste of excess paper and paint.
(Van Gogh inspiration #2)
Transitioning to even another form of art, artificial intelligence has become more advanced in the creation of music. Computers in the past have been able to play songs based on codes that humans input, but artificial intelligence takes this to a whole new level. AI allows computers to take inputted songs and then create similar, but new, music melodies. AI is not advanced enough yet to create words with these stimulated new melodies, but I presume this could change in the future as advances are made.
A great example of AI creating music comes from Sony’s Computer Science Laboratory. Last year the company created Flow Machines, an algorithmic program which creates computer-generated music. What’s really interesting is that the programmers were able to input different Beatles’ sheet music, and then Flow Machines generated by itself an entirely new melody that sounded similar to Beatles’ songs. Humans added lyrics over the new melody, which they named “Daddy’s Car.” What’s even cooler is that “Daddy’s Car” was named a semifinalist of the 2016 ISC (International Songwriter’s Competition). Even though humans were still required to create the lyrics over the melody, this AI technology is yet another example of technology disrupting industries and replacing human talent.
Finally, another great example of artificial intelligence creating music is Magenta, which is Google’s artificial intelligence music program. Besides being able to create songs, Magenta has the ability for people to play a tune on a keyboard, and then compile a similar tune to play at the same time, resulting in a near perfect duet. I say ‘near perfect’ because Magenta was developed in June of last year, so it is relatively a new project and still needs to improve. Right now the music is created with few notes and a keyboard piano sound. As time progresses, it will be interesting to see just how complex AI can become when creating music.
But even with all these examples, can these computer generated pieces truly be considered art, and if they are, can they truly be considered art at the same caliber as artists like Van Gogh? At this current stage, humans are still needed, such as to input multiple images and create lyrics over melodies, but I am curious to see how computers will continue to compete against manmade art.
Art and emotion are typically seen as interconnected, where it takes a special talent and natural born gift to create works of art. Do these types of artificial intelligence give way to the future of computers developing emotions when they make decisions? Or will computers remain simply grounded in code and algorithms? I guess time will tell…