Kendrick Lamar’s “The Heart Pt 5”: How They Did It
By now, thousands of people are watching Kendrick Lamar’s new music video for “The Heart Part 5,” during which his face changes to portray different celebrities who are Black men under fire or heroes who have passed away. The technology used to create this effect is likely NVIDIA’s Face Generator AI, which is a new generative adversarial network called Alias-Free GAN.
So what is a GAN?
A Generative Adversarial Network, or GAN is a kind of deep learning AI that generates images and other media based on an underlying distribution of data. The data includes human faces, songs, cartoon characters, text, or movie posters. In this process, an algorithm is set up to perform a task that automatically discovers and learns the patterns in input data (such as Kendrick’s face) in such a way that the model can be used to generate or output new faces (ex. OJ Simpson, Will Smith, Kobe Bryant). The Alias-Free GAN fixed early issues with StyleGANs and you can learn more about it here:
Deep learning allowed me to use an algorithm to redraw an image of Kendrick Lamar using the stylistic elements of another image. The software can learn the details of a source photo, then combine the contents of this image with styles of different ones, effectively transferring the styles to the original image. StyleGANs (face generators or “deep fakes”) take this technology to the next level, into video and animation.
What makes the use of an Alias-Free GAN so ingenious in Kendrick Lamar’s video is how it represents what he is saying at a given moment. For example, when Lamar’s says, “Friends bipolar, grab you by your pockets
No option if you froze up, always play the offense” his face changes to Kanye West’s (who supposedly suffers from the disorder). Or when he says, “In the land where hurt people hurt more people / F*ck callin’ it culture” his face changes into Will Smith’s (re: Smith’s Oscars slap). Lamar also pays tribute to Kobe Bryant and Nipsey Hussle.
This video is not just about showing off new technology. Lamar is using it to convey an important message:
But I want you to want me too (I want, I want, I want, I want)
I want the hood to want me back (I want, I want, I want, I want)
I want the hood
Look what I done for you (Look what I done for you)
Look what I done for you
Look what I done for you… and you cancel me? Scrutinize me or worse: murder me. Lamar lets us know that he is a Black man first and, as a rapper, he is telling you about what is happening to Black men in the public eye. He is also rapping about what happens when Black men become famous.
Should I feel resentful I didn’t see my full potential?
Should I feel regret about the good that I was into?
Everything is everything, this ain’t coincidental
I woke up that morning with more heart to give you
As I bleed through the speakers, feel my presence
It is because the video is so compelling that we soon forget that a deep learning AI is making things happen. We listen to the words, reflect, and talk to others about what the song and video means. Kendrick Lamar is creating space to discuss what is happening in his world that he shares with many others. It is amazing to see how AI was used represent his/their reality.