Personal Style: Where do personal style and AI intersect? *
Do we truly have our own sense of personal style, or are we just copying the influencers around us?
Investigating how well AI could copy my own personal style made me think of the nature of personal style. What does that mean in the era of large datasets that seem to contain all styles. Do we somehow just pick and choose from what is out there? What distinctive, original content lies inside of us? Is there nothing new under the sun?
People can slightly unleash themselves from the culture and community of influencers. We are distinct because we can surprisingly change our style. Machine learning that looks at past datasets can’t predict our shedding of old rules and taking in new ones. In addition, people can be exposed to events or engage in experiences that modify their personal style. We can possibly work with the world of AI by curating your own datasets.
Is it good to be guided by external forces acting on our imagination and actions? Maybe. Rules guiding creative behavior allow regular people who have something to say (don’t we all?) to use rules as adjuncts to their innate talents and educated abilities. Recognizing rules related to art and music can open up a wonderful dimension of human potential. Human expression is empowered when everyone is asked to be artistic by those who establish their community’s intellectual ecology. These mentors can encourage silent voices to add alluring qualities and expressive communication to the world around them.
Reducing Creative Expression to a
Flow Chart…Yuck
Everyone is creative. However, expressing creativity outside the daily requirements of problem solving and moving into art and music goes through three basic steps.
- Imagination of creator
- Technology to implement creation
- Physiology of viewer
But there is feedback, whether be from how your paint is showing on the canvas or the music to your ears or the feedback from other people (Picasso and Braque are a famous example of productive peer critiques).
It could be modeled as follows:
A picture of creative expression
Employing rules to bring personal mental images or musical creations to an audience broadens access to these processes. Seeking to communicate creative impulses with rules and structure extracted from cultural norms and other sources democratized artistic and musical expression. Children learn normative rules in speaking and writing that quickly let them use these abilities to express individual thoughts. A parent understands, “I miss daddy” as well as a lengthy narrative about loneliness and familial connections. A child’s stick figure drawing showing two people holding hands, presents as clear a meaning as a painting of the same thing.
We can quickly use normative rules to engage in art forms that may have traditionally been moderated by human gatekeepers. Introducing these rules can open a corridor for richer expression of ideas. Moreover, coupling these rules with intuitive software, online instruction, and the unfettered communications on the internet, allows more inclusive entrance into creating art and music.
Rules can guide the novice. For example, in music the relationship between notes that make up a chord are established and culturally conditioned. Moreover, the progression of these chords follows commonly recognized sequences. Therefore, one can couple lyrics with a chord progression that turns poetry into music. The visual arts also have rules related to guiding compositions and color. Moreover, artificial intelligence has been used in music and the visual arts to identify and redeploy patterns in new work.
I’m Offended!
Musical composers should be offended at the notion of creating music by simply accompanying lyrics with a rule-based (or sampled) approach. This offends the depth, breadth, and creative majesty of musical composition. A traditionally trained artist will likewise be offended at artwork produced by morphing existing images and applying colors from a standardized color scheme. However, musical and artistic creation can be democratized by identifying patterns and normative rules which are redeployed in a new, personal form.
One Last Thing: People can create surprise
Surprise is an idea that fascinates me and I have written about it in an early blog post. The ability to surprise is a distinctive human attribute that is a challenging notion for AI. However, surprise in one’s environment may be threatening. Both behavioral and neural responses may be rooted in minimizing surprise. Surprise is the inverse of the probability of observing the outcome. People’s expectations are moderated by experience or education to reduce surprise. A baby is surprised by snow falling from the sky, the surprise soon goes away. However, surprise can be enjoyable because it upsets the norm. Too much surprise is unappealing or even frightening, while too little is boring.
Surprise falls between the design concepts of unity (or familiarity/expectations) and variety (or novelty/tension). Unity refers to the manner in which different aspects of art combine to form a sense of wholeness. This is achieved by such means as repetition of visual elements in artwork or audible elements in music. Variety is the opposite. It disrupts repetition by including changing elements. People enjoy variety, such as watching people’s faces. They are symmetrical left to right but not top to bottom; however, they relate in a pleasing way. There is a rhythm to facial movement. In visual art, variety can be the inclusion of different colors, shapes, and textures. In music, variety can be using notes outside the key or by syncopation.
Music and art have many examples of surprise, and these normally carry the distinguishing names of artistic movements, from Impressionism to Jazz. The rules within a tradition are challenged by such works as Marcel Duchamp’s 1917 Fountain and Igor Stravinsky’s 1913 Rite of Spring. These were overt affronts to the status quo.
Here is a remix of the song with a discussion of process. The remix is at the end.
* Much of the last two blog posts and YouTube song derive from my most recent ASEE paper, Engineering Art: Democratizing artistic expression using normative rules. The proceedings have not been published yet.
