Because the ubiquity of generative Artificial Intelligence (AI) conversations is as infuriating as it is essential, we once again find ourselves discussing its impact on art. Apologies to everyone who’s overwhelmed. This is our life for the foreseeable future. Best we could advise is continual receptiveness to questions about the 21st century’s greatest cognitive disruptor. And in the instance of this essay, our questions are with regard to Fela’s weapon of choice: music—specifically, Nigerian music.
Hip-hop icon and one of the architects of Afropop (Afrobeats), eLDee tha Don, released a video on Instagram on Wednesday, 22 July, 2026, previewing I Wish You Well, a new song on his forthcoming project. Reactions were mostly positive, many excited at the prospect of new music from the man who had a stellar career as part of the Da Trybe collective, the Trybesmen triumvirate (with rappers Freestyle and KB), and as a solo act; especially since this would be his first project since 2012’s Undeniable.
The video broke containment to X the following day, and much unlike Instagram, reactions were divided. On one end, eager fans celebrated the possibility of new eLDee. On the other end, music professionals—particularly, producers—called out perceived use of generative music AI. Highlighting certain tell-tale features common to songs generated by Suno AI, most argued against it being used by someone of his status. The conversation then devolved into a maelstrom of ad-hominems targeted at supposedly unknowledgeable Gen-Zs, among other reductive arguments.
He wouldn’t be the first OG artist to be accused of using AI in recent times. In June, Peter Okoye, or ‘Mr P’ as he’s also known, came under fire for the use of AI on his comeback single, I Can’t Look Away. But whereas the defence then was that AI use didn’t matter as much, in eLDee’s case, advocates instead question the idea that an artist of his calibre would defer to AI to make ‘lyrical’ music. The assumption is that his skill level erases any likelihood of employing that sort of outside help.
However, there’s enough to suggest that the Don indeed utilised gen AI. He’s employed AI for a host of other functions across his social media. Scrolling through more recent Instagram videos, you’ll find: an AI-generated thumbnail for his podcast, Nigerian American, with that font from hell, and pictures of himself, The Beatles, Bob Marley, and Usher; an AI-generated audio script and even a video-script straight out of the uncanny valley; and a video of the rapper in full Samurai armour, among others. Clearly, he’s adept at optimising gen AI tools. And clearly, from the video podcast being uploaded to DSPs, he’s also proficient enough to work his way around edits and what not.
Beyond previous use, there’s the nature of the recording itself. When asked to describe specific features indicative of gen music AI, most of the producers deferred to intuition and familiarity with such production. Some outlined the polished nature of the backing vocals, as well as the rhythm of the verses as a dead giveaway. Unlike writing (rule of threes, metaphor abuse, excessive imagery etc) and videos (abundant artefacts, poor continuity etc), gen AI music isn’t as widespread, and hasn’t been engaged with as much, for most hobbyists to pick out distinct features. And that’s without accounting for the numerous false identifiers—like the em-dash. However, when you’ve spent enough time listening to such production, and even listened through tutorials on ‘how,’ as we did while working on the story “When is AI Music Not Large Language Mediocrity?” you notice certain tics lending credence to this hypothesis.
eLDee’s subtle country music inflection at the beginning of the video is one. As are the rhythm and beat-count of the lyrics, bar the hook. The arrangement for the “Next day she hit my phone, she need a loan” section is another tell. eLDee hitting an unnaturally high note just before the hook. As had been pointed out earlier, background vocals also sounded strangely polished. Additionally suspect, but not as confirmatory, are the rhyme scheme and pulsating piano chords, which could be anything or nothing. For the former, these self-appointed long-time fans won’t tell you that dating as far back as 2001’s Oya with Da Trybe right up to the Sarz-produced Joor oh recreation, Wash Wash, in 2012, his delivery always had a distinct Nigerianness to it. Even at his most American —think autotune dancehall magic on Been There, Done That—eLDee didn’t lose that flair. Up until the hook, most of that twang is absent.
Of course, none of this is absolute proof. This writer’s summation could be just that—mere, unfounded summation. And if there’s anything the literature witch hunts have taught us, it’s that accusations of this nature can quickly take on more malevolent tones. The video which went up afterwards shows the song being mixed, separated vocal stems and all, with eLDee laughing along with folks decrying the AI accusations in the comments section. That’s going to be the extent of any complaints. And honestly, there’s no reason why else. If he did use AI to bolster production and songwriting? Fair. If he didn’t? Also fair. There’s little to gain from hounding an artist of his stature, whose possible AI use isn’t particularly as harmful as, say, unauthorised covers of existing music. At worst, it’s more of an ethical concern regarding how Suno or whatever tool, sourced their training data.
It’s also important to consider why an individual of his stature would return to music after so many years. Why a peer like Ikechukwu released Triumph Over Tribulations in June, an update to his 2024 album, O.G.Uncle.Killz, which was also his first project since 2009’s The Alliance Reconstructed. Why producer ID Cabassa took on the landmark Unfinished Business album last year (We won’t be forgetting about Odumodublvck’s Olufunmi shocker verse anytime soon). And why this year, the Coded Tunez magician released the two-track Affairs of the State EP—over-saturated GPT-generated album cover in suit. It’s pretty simple when you think about it: they’re creators. They got into music because they loved making music. And as anyone afflicted with the curse of the muse will tell you—including 92-year-old Prof. Wọlé Soyinka—she can be pretty hard to shake off.
In a 8 May, 2026 video where he announced the album, eLDee admitted this much, stating that he’d been fighting the urge to return to music and was now motivated to do so, but with the caveat that the music creation process would be interactive. Also, responding to Culture Custodian’s video commentary on the excessive use of lamba in current Afropop on 15 July, the rapper mentioned being told to include more lamba in recording new music, a situation he considers to be detrimental to the genre. He’s tapped into all that’s going on, quite all right. It’s not so much a desire to jump on trends and ride the wave of success as it is a desire to create. And as he’d go on to rap in a 28 May freestyle, personifying hip-hop like Common on the 1994 classic I Used to Love H.E.R, music is his love.
Unfortunately, audiences aren’t so receptive to these older acts. Some 2020 stars haven’t even regained the spotlight, talk more of these OGs. Music is unforgiving in its ageism. Our industry has also failed to develop the infrastructure that would encourage touring or second-winds for the folks who paved the way. This is why malicious accusations of fame-chasing fame don’t hold up. Not when they’ve moved on with different careers, some almost 15 years deep into oversight roles. But leave it up to the mob to hurl as many boulders as possible in their search for drags.
Do we exempt older artists from the crusade against ethical gen music AI use? It’s an easy exemption if you do not flip the script, if you do not sit with how this music emerges. For all we know, older artists might be creating music from a model trained on their early, minimally protected work. Would it then be ethical if a younger rapper utilised AI to produce lyrics with the swagger of Big Boy or the melody of Bosi Gbanga? Not a question you’d want to answer, in all honesty.
As the conversation about eLDee raged on X on Thursday, producer Kid-Konnect contributed his perspective on the subject, comparing the inevitability of AI to his adoption of music production software in the early 2000s, when many insisted on hardware only. This comparison doesn’t hold up under scrutiny. But it’s understandable why creatives from that generation and above—legendary Hollywood directors George Lucas and Martin Scorsese, inclusive—don’t see another outcome. It’s younger generations of creatives who are vocal about gen AI adoption, although, even they have begun incorporating this technology on a large scale.
If all else fails, quality will still remain as a divider. Without resorting to malicious mudslinging, the bulk of AI music that exists is fifty shades of disgusting. Crappy lyrics. Crappy production. Unsettling vocals. Syrupy hooks and unnecessary baptisms of crowd vocals. As bad as the current standards are, rest assured AI-assisted records won’t automatically surge atop TurnTable Charts or DSP charts, not even with the greatest stream farming machinery. Let the uncs and aunts do their thing. Allow them to breathe. The OGs call you fans, bear no hand no in their dragging.
0 Comments
Add your own hot takes