Teenager Teaches A.I. to Rap Like Kanye

March 19, 2017, 8:28 PM UTC

A 17-year-old from West Virginia has used an archive of Kanye West lyrics to train a neural network to write rhymes on its own. The results are braggadocious, intermittently obscene, frequently incomprehensible, and laced with nonsensical name-drops.

In other words, they’re pretty good rap lyrics.

Get Data Sheet, Fortune’s technology newsletter.

The rapping robot was created by Robbie Barrat, who told Quartz that he put it together in a week, at the urging of his high school programming club, using the open-source software PyBrain and a Linux laptop. Most impressive of all, this is apparently Barrat’s first working AI.

Here’s one choice sequence of Barrat’s creation going to work:


I’ll touch every curve of your favorite author

No more wasting time, you can’t roam without Caesar

Back when Gucci was the best summer ever

Before Cam got the hundred with the peer pressure


It’s not exactly Wallace Stevens—it’s hard to find much in the way of consistent themes or narratives in the robot’s work. But the rhymes definitely echoes West at his most wildly free-associative. (And for the curious, “Cam” might be Cam’ron, a now sadly overlooked rapper Kanye collaborated with back in 2007).

Barrat has posted a couple of videos highlighting his creation’s lyrics, though he’s quick to point out the A.I. didn’t make the backing tracks. And the voice generator he’s using is pretty short on flow, charisma, and even comprehensibility—so far, Barrat’s focus is clearly on words.

As Quartz points out, this is about more than just one quirky conceptual project. A.I. is a hot field right now, crucial to innovations from driverless cars to image-based search. While it’s not ready to put human performers out of work, Barrat’s project is directly relevant to the huge field of natural language processing, ala Siri and Amazon’s Echo.

The fact that he did it using free programs, far away from U.S. centers of tech innovation, should give us all a little extra hope for the future.

Read More

Artificial IntelligenceCryptocurrencyMetaverseCybersecurityTech Forward