028: Existential Threats
Friends,
Happy Monday, and for the UK folks, happy Bank Holiday.
I can only apologise for an extended period of inactivity. It wasn’t in my plans to not publish any work for circa four months, but I appreciate the fact that subscriptions and viewership have not taken a hit as a result of that. I am still striving to find the perfect balance between doing something cool on the side of juggling day-to-day commitments, and I am truly humbled that you all stick around regardless.
But anyway, let’s get back to business…
Artificial intelligence is the buzzword of 2023. In the world of hip-hop, it became a hot topic and cause for concern when viral posts showed the extent of AI’s ability to manipulate artists’ voices to create fake covers of tracks, or even in some instances, produce “original songs”. What then followed was discourse around the state of play in music and other art forms, whether it was a sign of the end, or a simple shrug of the shoulders. The answer lies somewhere in the middle, with overlaps from the AI/hip-hop subject matter into other disciplines concerning this vast Venn diagram.
Hip-hop turned 50 years old this month, which means I’ve been an obsessive fan and follower for over half of it’s existence. I recall when I had the pleasure of interviewing the guys over at My Type Of Hype last year, where the OG E-Major said “Hip Hop is a culture, not just a genre, it is 50 years old which means it has history”. It’s with that in mind that I felt compelled to think out loud on this topic of AI/hip-hop, and try to make sense of what it really means for hip-hop now and in future. As somebody who happens to work in tech, where my current day-to-day involves being part of a team building an AI/machine learning product, my curiosity between the two subjects peaked. But as I mulled over my thoughts, I also realised that this isn’t the first time hip-hop has faced threats to its existence.
With AI, in short; I don’t think hip-hop or the art of rap need worry about being “replaced” by machines, as is the case with music and the arts overall. I think most, if not all efforts for this will be challenged, rejected or deemed null. Ultimately, I believe this is just another entry into the pile of get-rich-quick - or in this age, get-attention-quick - schemes of culture vultures, cosplayers and manipulators of hot trends that wish to initialise their first steps into 15 minutes (if that) of fame. Using the viral tweet I linked above in the opening paragraph is a prime example. One quick advanced search on X from the account @rpnickson showed him mention “rap” or “hip-hop” less than 5 times in total over the entirety of his time tweeting ever. As for “music”, that only came up alongside web3 and NFT subjects on a handful of occasions. Yet he suddenly goes viral when “The music industry is forever changed”, using one of the most popular hip-hop artists from past and present to deliver his example and farm interactions. Surface level shit by a grifter from top to bottom if I’ve ever seen one.
Furthermore, another example came in the form of a Drake and The Weeknd deepfake collab, which journalist Joe Cocarelli discussed and broke down in his interview with The New York Times; a very interesting listen (or read, should you prefer the transcript). Joe and I are on the same page, but I’d like to scratch this surface a little bit deeper with some tech context too. A problem we have at work with our AI/ML project is data, and specifically the amount of rich, real data we have at our disposal to build models from. An AI model to make music, or a genre of music specifically, would arguably have to use the entirety of a genre’s historical collection in order to “predict” and “learn” how to make a song of that genre or style in question to stand any chance of producing notable results. These models need data sets in the tens of thousands and above to work well by most reasonable measures. But with every artist having their own unique perspective, its hard to say if the model would spit out something streamlined, and effectively, anything of substance or quality. In the case of hip-hop, it would hypothetically be extremely scattered and disjointed, with so many different styles, cadences, personalities and other traits displayed by rappers in their music.
In addition, it would be even more difficult to do this with a specific artist - lets say Jay-Z - because there isn’t enough data. Jay-Z has 13 albums plus features, collabs and remixes. We’re talking in the region of perhaps 200-300 or so pieces of data to use and reference. It sounds like a lot, but in the world of AI/ML, its not enough to build something sturdy or strong enough to perform to some level of expectation. Even sticking with the Kanye West example from our aforementioned viral post, would a model based on his first 3 albums have predicted the drastic shift in sound and direction from him to create 808s & Heartbreak? Edge cases are always the most difficult to capture. These things would prove even more challenging to artists with smaller discographies and catalogues. Human creativity and unpredictability still wields its influence in these instances.
“But the question AI raises is, do we even need that connection?
Do we just want something that sounds pleasant enough in the background? Oh, that sounds like Drake, or that sounds like Bob Dylan. Or do we need to know that this is coming from the depths of their soul and from their lungs and their heart? Maybe we do or maybe we don’t. But as AI keeps coming in music and in other art forms, we’re all going to have to answer that question for ourselves.”
- Joe Cocarelli, NYT Music Culture Reporter
There’s also financial elements at play here too, and unfortunately, cost-cutting can cloud the judgements far and wide, even with our very own pioneers and legends. In essence, we’ve seen many onslaughts and threats to hip-hop’s purity before, internally and externally. Listeners need to take a stance with what they want too. I’ve mentioned in previous articles that consumers share an equal burden in terms of the state of play. There is, of course, some cause for concern for artistic quality and integrity, but some might argue that this ship has long sailed with no turning back. Is the entrance of AI, tech/tech bro culture and any associated expansions a by-product of hip-hop’s barriers being down for too long? Is this even the main thing hip-hop artists and the community needs to fear the most, considering the long standing differences between musicians and their labels? There are certainly some self-inflicted wounds, but there’s a lot more fish to fry than what we’ve seen in 2023. AI might need to wait in line.
Equally, it’s not coincidence to see this AI issue propped up in what can only be categorised as a fast-food era, with troubles brewing away from beyond the low vibration music as well. Journalism, fandom and everything else seems to be lost in some social media infused, gossip-ragged battle for attention and its becoming tiresome to say the least. The focus has quite clearly shifted from the art to the product. Visuals, cults and shock factor define the trajectory more than bars, beats and message. Typical, leach-like, opportunistic behaviour in Silicon Valley fashion. Nothing to see here, perhaps.
But fear not friends, because despite these threats, I’m optimistic that neither of them have enough to materialise into a sustainable enemy. Moreso, the points raised here are designed more to make yourselves think. To think twice about headlines, to think twice about the town criers spreading these updates and trends, to take a step back and assess what’s in front of you. Hip-hop has always taught us to stay woke (the real definition), to use our third eye, and to cut the grass low and expose the snakes. I’m fairly confident everything will be alright, so long as we exercise our judgement and remember the bars of wisdom from our favourite philosophers.
Before I sign off, here’s some recommended reading and viewing on the topics discussed in this article:
https://www.morningbrew.com/daily/stories/2023/07/21/the-true-story-of-oppenheimer?mbcid=32155714.1344693&mid=2d0e95eb814131e5b294f2e1c2424b1d&utm_campaign=mb&utm_medium=newsletter&utm_source=morning_brew
https://www.morningbrew.com/daily/stories/2023/07/21/chatgpt-seems-to-be-getting-dumber?mbcid=32159845.1186422&mid=2d0e95eb814131e5b294f2e1c2424b1d&utm_campaign=mb&utm_medium=newsletter&utm_source=morning_brew
https://www.morningbrew.com/daily/stories/2023/05/16/ai-is-in-its-regulation-era?cid=31505854.1212391&mid=2d0e95eb814131e5b294f2e1c2424b1d&utm_campaign=mb&utm_medium=newsletter&utm_source=morning_brew
https://www.wired.com/2014/01/geoffrey-hinton-deep-learning/?utm_brand=wired&mbid=social_twitter&utm_social-type=owned&utm_source=twitter&utm_medium=social
https://www.thenation.com/article/economy/artificial-intelligence-silicon-valley/
https://www.ft.com/content/7ca8e38d-8ecf-4de2-870d-9bd0a2476147
https://hiphopdx.com/news/blu-exile-below-heavens-album-no-money-anniversary
https://hiphopdx.com/news/nas-older-rappers-follow-blueprint
https://www.passionweiss.com/2022/12/01/boldy-james-interview-mr-ten08-new-album/
https://www.okayplayer.com/originals/roc-marciano-reloaded-album.html
https://www.bbc.co.uk/iplayer/episodes/p0dj6ysm/fight-the-power-how-hip-hop-changed-the-world
Do yourself justice, utilise your mind.
WhosAria