Flying houses and cars, rocket-powered mail services, and widespread telepathy. These “Jetson-like” innovations represent just a few of the hilariously inaccurate predictions made in the ’70s about life in the 2020s. While most of the educated guesses our secular brothers and sisters made about contemporary life were miles off the mark, they weren’t wrong on one thing. In fact, their prediction has not only come true, but has become one of the greatest threats to all forms of original art today: artificial intelligence.
No, Ex Machina’s robots haven’t started curating art collections yet. However, the recent emergence of AI-generated artwork and music has rippled through both industries.
Thanks to TikTok, Wombo Dream (available for iOS, Android) has become one of the most accessible forms of AI artwork. By simply typing in phrases or keywords, Wombo uses AI to generate artwork that combines the word prompts with artistic murals existing images. In a matter of seconds, Wombo’s AI is able to stitch together intricacies in art that would take human artists hours or even days to create, even with specifically vague prompts like “Galactic Archeology With Metal-Poor Stars.”
With billions of images available in an instant to an advanced AI like the one behind Wombo, the sophistication of the AI-generated artworks is startling, but nowhere near as alarming as the music they can generate.
The extensive capabilities of AI music encompass a comprehensive set of musical processes, including composition, performance, digital sound processing, and even interactive composition. Many websites out there can emulate something similar to what Wombo presents and offer an AI capable of producing millions of songs based on user specifications. But music AI is far more intuitive, having the ability to react to a live performer in real-time. Used in this way, AI can replace entire live bands and orchestras, producing the same quality of music in less time, with less confusion and more harmony.
While less advanced AIs use internet databases to power their machine learning, music AIs use neural networks to mimic how the brain works when creating music. When you throw pieces of music at these AIs, they essentially learn its patterns and frequencies by being repeatedly exposed to it.
Perhaps one of the most disturbing examples of this technology is its use by one of the most musically disadvantaged fanbases in rap music: Playboi Carti fans. A simple YouTube search yields many AI-generated Carti tracks made using this technology. Fans even created an EP for Carti with AI called DIGITAL BUTTERFLIES. The project uses Carti’s real voice, famous ad-libs, and even frequent Carti collaborator Pierre Bourne’s sound kit to create a lively, psychedelic six-track trap project that’s almost as polished as something Carti himself did created at the beginning of his career.
*embed “digital butterflies EP on youtube*
Aside from being downright creepy and soulless, this clearly presents a plethora of pressing problems for the music industry. From further blurring the already murky lines regarding posthumous music to opening up even more avenues for artist exploitation to can and will the mere presence of AI music in its current state be an obstacle.
One way or another, original art is becoming scarcer, whether we like it or not. Much like how we consume social media, art needs to be viewed through the eye of a skeptic.
Luke Modugno is a digital editor. Follow him on Twitter: @lmodugno5.