Can Machines Be Creative? Meet 9 AI ‘Artists’

3 min read
Curated from livescience.com →

One of the behaviors considered to be uniquely human is our creativity. While many animal species create visually stunning displays or constructions — think of a spider’s delicate web or the colorful, intricate structures built by bowerbirds — they are typically created with a practical purpose in mind, such as snagging prey or seducing a mate.

Humans, however, make art for its own sake, as a form of personal expression. And as computer engineers attempt to imbue artificial intelligence (AI) with humanlike capabilities and behaviors, a question arises: Can AI create art?

The AMC series “Humans,” which returns June 5 for its third season, is populated by Synths — intelligent robots that resemble people, save for their unnaturally green eyes. At the end of the last season, millions of Synths had “awakened” to consciousness. Plenty of humans are unhappy about that, presenting Synths with more immediate problems than getting in touch with their inner artists. Nevertheless, exploring and indulging in creativity is introduced in the show as one of the avenues that is newly open to Synths, now that they are self-aware. [5 Intriguing Uses for Artificial Intelligence (That Aren’t Killer Robots)]

Even in the real world, AI has demonstrated that it can perform feats that are unexpectedly creative. Under the guidance of programmers, different types of AI have produced original songs, paintings and digital artwork; some bust out their own dance moves, and others spit rhymes like Kanye.

Here are just a few examples of some of today’s “artistic” AI. Perhaps the Synths will express themselves in similar ways …

In September 2016, a human-AI collaboration produced a cheery pop song that sounded remarkably like a tune composed by The Beatles in the 1960s. While the lyrics were written by a person, the melody sprang from the “brain” of a computer. AI used a system called Flow Machine to incorporate data from 13,000 lead sheets — musical notations for the elements in a pop song’s melody — and produce a composition that the programming team dubbed “Daddy’s Car.”

Care to dance? Inside a dome framework covered by white fabric, AI recorded and processed dance moves performed by people. It then incorporated those dance steps into a projection of a virtual dancer, which performed a sequence of moves that the AI “remembered” from the human dancers. However, the virtual dancer also matched its rhythm or style to the most recent performer in the dome, turning every dance into a collaboration.

Algorithms can teach AI about the basic elements of a visual style so that neural networks can apply them in original ways to create unique and intriguing artwork. In one example, an AI created “botanical dinosaurs” — images of Triceratops, T. rex, Stegosaurus and other dinosaurs, composed entirely of plants and flowers. This demonstrated a technique called style transfer, which incorporated an art style — in this case, botanical illustrations — and rendered it in the shapes of dinosaur bodies.

Can a computer know what scares you? This one can. An AI project aptly named “Nightmare Machine” revealed what makes pictures frightening by using deep learning, a system of programs and data structures that form connections in a manner similar to that of neurons firing in the human brain.

Continue Reading

Enjoyed this summary? Read the complete article at the source:

Continue at livescience.com →

Yves Mulkers

Yves Mulkers is the founder of 7wData and a widely followed voice in the data and AI community. He curates the 7wData and AI Beat newsletters, reaching hundreds of thousands of data and AI professionals, and writes on data strategy, analytics, AI, and the evolving data ecosystem.