AI’s Big Three Crown a Victor in the Drake vs. Kendrick Beef
Dr. Newton Campbell — Computer Scientist, Hip-Hop Fan, Space/Climate Advocate, Pragmatic Goody Two-Shoes
In the world of hip-hop, few feuds have been as highly anticipated and deeply analyzed as the seemingly concluded lyrical battle between Kendrick Lamar and Drake. For years, fans have debated who truly reigned supreme in their thinly-veiled disses and subliminal shots traded through verse.
But now, an allegedly impartial panel of judges has rendered their final decision.
The AIs Weigh In
The judges? AI’s current Big Three. A council of the latest advanced artificial intelligence unanimously scored the bout after meticulously analyzing the rappers’ full lyrics, wordplay, storytelling ability, impact, influence, and every other facet of their artistry.
Our judges consisted of:
- ChatGPT—GPT-4o is a new AI model from OpenAI that can understand and respond to a combination of text, audio, images, and videos. It excels at understanding vision and audio inputs and is especially good at recognizing speech in multiple languages.
- Meta AI — Built on Llama 3, Meta’s (formerly Facebook) powerful large language model is capable of analyzing longtexts, following complex prompts, and drawing insights from massive datasets.
- Gemini — An ambitious open-source project from Google aiming to create one of the world’s most capable AI language models by crowdsourcing training data from thousands of contributors globally and supporting larger inputs.
Each offers AI chatbot services that are affordable and accessible, with different capabilities available to free tier and to Plus users with higher message limits and some additional features.
These AIs offered an interesting analysis of the rap beef void of industry affiliations. They judged strictly based on the merits of the lyrics, flows, themes, production credits, and context surrounding both artists’ releases over the course of the beef, starting with First Person Shooter.
With no profit motives, corporate ties, or fan allegiances to sway their judgment, the AIs offered a purely computational analysis of both rappers’ bodies of work to conclusively declare one as the superior wordsmith and musical visionary in this high-stakes beef.
The Verdict — Unanimous Decision
And the winner, by unanimous decision is:
Someone go get him a Grammy right now.
After careful consideration of the lyrical content, flow, impact, and overall artistry displayed in the diss tracks exchanged between Kendrick Lamar and Drake, all three AI judges independently concluded that Kendrick Lamar emerged as the victor in this rap battle. Their analysis and reasoning was the interesting part. Let’s take a look at where the judges agreed and disagreed.
Let’s start with the final verdict of ChatGPT:
ChatGPT with GPT-4o highlighted Kung Fu Kenny’s “consistent focus on deeper cultural and personal critiques, blending lyrical complexity with powerful messages.” The AI emphasized Lamar’s ability to undermine Drake’s credibility and public persona through introspective and reflective tracks, giving him a clear edge. It pointed out “6:16 in LA” and “Meet the Grahams” as the songs that really solidified Kendrick’s position in the rap beef. Note that it also had a bit of trouble with nuance, as it pointed out K-Dot’s “empathy for family members” during the beef, later citing “Meet the Grahams.”
For Drake, ChatGPT praised his variety of insults, likely honing in on the fact that so many were launched in so many different directions. It consistently described Drake’s verses as confident, detailed, and dismissive. However, by the time we got around to “The Heart Part 6,” ChatGPT noted that the 6 God’s focus was mostly shifted to defending his public persona rather than addressing deeper cultural critiques. And it handed Kendrick the win.
Next, let’s look at our judge from Meta AI:
Meta AI echoed ChatGPT’s sentiment, stating that Lamar “won” due to his lyrical precision, consistency, storytelling, and ability to take the moral high ground. The AI also noted Lamar’s strategic timing in releasing his diss tracks, allowing him to control the narrative. Most human judges have pointed this out as the reason he won the beef.
Meta AI’s judgment was fairly against Drake for most of the beef. It constantly called Drizzy’s attempts commercially appealing and catchy but lacking the lyrical depth and consistency to effectively counter Kendrick’s narrative. It maintained that critique from Push Ups all the way through “The Heart Part 6."
And finally, the judgment from Google’s crown jewel Gemini:
While acknowledging Drake’s commercial success and emotional vulnerability, Gemini ultimately sided with K-Dot, often citing his lyrical complexity, impactful diss tracks, and cultural currency as the deciding factors. While it notes that Drake has moments of brilliance, his reliance on unproven personal attacks and certain repetitive themes ultimately undermine his own efforts. Kendrick’s willingness to address social issues and explore deeper themes solidifies his position as the winner of this lyrical battle. You can also tell it dug into the past in its reasoning, as it brings up past diss records like Control. It also gave a great disclaimer at the end, which we need to talk about.
Are we Battling Ghosts or AI?
The Kendrick Lamar vs. Drake beef, as judged by The Big Three AI models, offers a fascinating glimpse into artificial intelligence's current limitations in understanding and evaluating creative expression. While the AI models’ consensus on Kendrick Lamar’s victory might seem definitive, it’s crucial to remember that these are just tools, not infallible judges.
Each AI model, with its unique training data and algorithms, brings a different perspective to the analysis. They gave very detailed feedback as I introduced them to each song of the beef. ChatGPT, with its vast knowledge base, focused on lyrical complexity and cultural impact. Meta AI emphasized storytelling and strategic timing. And Gemini highlighted the importance of lyrical precision, consistency, and culture.
However, these models also have their limitations. In many responses, they struggled to grasp the nuances of cultural context, humor, or sarcasm, potentially leading to misinterpretations. They might also have been influenced by the popularity or prevalence of certain artists or styles, overlooking the merits of less mainstream voices. At the end of the day, they are feeding off of a massively biased dataset: The Internet. Try as the companies that own them may, the data these models are trained on carries all of the capitalism, racism, sexism, homophobia, and xenophobia of our collective societies. So proceed with your own analysis with caution.
Also, let’s talk science: While the AI models provide interesting insights into the Drake vs. Kendrick beef, it’s important to note that this analysis is based on a single run of each model. To draw more robust conclusions, we need to remember that most AI models are just sampling their answers from random probability distributions:
When AI from the Big Three gives you an answer, they find a point along a probability curve and form an answer based on that point. Think of each model as a huge jar filled with different colored jelly beans. Some colors taste better than others, so you’ve added more of those to the jar. Each jelly bean represents a possible answer to a question. When you ask the AIs a question, it’s like they’re reaching into the jar without looking. They’re more likely to pick a “good-tasting” jelly bean because there are more of them, but it’s still possible for them to grab a less tasty one. Furthermore, each “good-tasting” jellybean has a slightly different size and shape. Each of the Big Three’s AI models is trained on a massive amount of data from the Internet. So when you ask them a question, there are lots of possible right answers they can choose from. And they use what they’ve learned to try and pick the best answer. So each time you ask, the response will be slightly different, imperfect, and inconsistent. Because there are a heck of a lot of possible jellybeans.
An appropriate judging experiment for this beef would involve running these queries thousands of times to account for the inherent randomness in AI language model generation. And phrasing their queries in thousands of ways. This would help to assess the consistency of the models’ responses and identify any potential biases or limitations in their analysis.
Therefore, while the AI-generated analysis is insightful, it should be interpreted with caution and considered alongside human interpretation and analysis. By acknowledging the limitations of AI and incorporating diverse perspectives, we can gain a more comprehensive understanding of complex cultural phenomena like rap beefs.
In Conclusion
While human opinions on the G.O.A.T. debate will surely rage on, an AI consensus has now been calculated. The machines have analyzed the receipts, run the numbers, and declared Kendrick Lamar the victor of this particular feud.
Rap battles, like any art form, are inherently subjective. AI models, while capable of analyzing patterns and structures, may not fully grasp the emotional impact or cultural significance of certain lyrics.
Therefore, while the AI verdict is intriguing, it’s essential to remember that it’s just one perspective. Human interpretation and analysis remain crucial in understanding the complexities and nuances of creative expression. The AI’s analysis can serve as a starting point for further discussion and debate, but they’re Not Like Us, and should never have the final word.
But hey, it’s beef season in hip-hop, and whether you agree with the AI or not, one thing’s for sure: we’re all winners when it comes to great music. So let’s crank up the volume, enjoy the lyrical warfare, and hope for more bangers to drop this summer. After all, what’s hip-hop without the competition? Just remember to keep it on wax and appreciate the artistry, no matter which side you’re on.
Try the prompts for the chatbots mentioned in this article here.
