đ” Burner accounts, AI slop and a fork in the road moment for music marketing
Is it time to stop and rethink how we're strategising marketing?
Hi there -
If you arenât yet familiar with the term âAI slopâ, allow me to enlighten you. After all, Iâm quite confident youâre going to be hearing about it a lot soon.
AI slop is the term for the endless torrent of AI-generated imagery (for the most part, though it can manifest in other forms of media too) that is now flooding social media. Facebook appears to be ground zero, but Twitter/X is also drowning in it, and even Threads has shown signs of falling foul as the US election nears.
The recent Hurricane Helene was a case in point; thousands of images swamped social media of things like a young girl clutching a puppy, sitting in a boat amid a flood (pic shown further down this email). The problem is of course that the image is entirely fake, and yet this doesnât seem to stop it getting absurd levels of engagement - though one would be wise to even question whether the reactions are real, or just other AI bots going about their business. The zombie internet may be less of a theory with every passing day.
The end result is the overall quality of social media feeds moving into a persistent decline. Make no mistake: itâs all getting more and more shitty as AI desperately seeks to prove its worth by, ironically, overwhelming us with dross. It also has one wondering whether any metrics are even reliable now. How many of those views, likes or even comments were actually created by humans?
I was contemplating the AI slop issue earlier when an email arrived in my inbox, showing a case study from a social media agency, illustrating how theyâd used burner fan accounts (i.e. fake accounts, designed exclusively to generate fake fan content) across social media to drive exposure of an artist.
For me, thereâs an intersection on a venn diagram between AI slop and this kind of fake fan content. As the decline in content quality continues, Iâd think weâre not far away from AI-generated fan content flooding out there to try and drive awareness of an artist. Frankly, I think it quite likely weâre there already; agencies just might not be prepared to own up to it yet.
Burner accounts somehow feel a little like bought video view counts, in that both are ascribing value to something that might not have any value whatsoever. Did buying millions of fake views really help an artist? Does spamming video content across TikTok or Instagram, as if youâre a fan, help either? Are those views all genuine? Whatâs to stop the agency running their own AI bot network just watching this content and engaging with it?
Iâve no doubt a social agency might pop up at this point to tell me I am wrong, and that getting thousands of views on content all helps, but I could only respond by asking them to prove what it actually did for the artist beyond just provided very, very fleeting engagement metrics that they might struggle to prove are even legitimate.
Ultimately, this is the landscape we have collectively bought into, so it is tough to blame social agencies for making money from acting in this manner. However I think it is yet another case for questioning whether any of these approaches really amount to effective, useful marketing that contributes to long-term connection and growth, and with that a consequent return on investment.
The music industry badly needs to stop and review how budget is being committed on artist campaigns, asking what materially delivers value and what does not.
Donât get me wrong: great, engaging content from artists can resonate. However spamming âuser-generatedâ material from fake accounts just feels desperate in the extreme, symptomatic of how much the tail is now wagging the dog in terms of our collective use of social media platforms.
Personally I feel we are long overdue a review in which companies step back and really work to examine how success is measured, which data is reliable and whether money is therefore being apportioned into the right places.
We can all do better, and as the quality of content on the internet, especially the social internet, moves into something of an AI-fuelled death spiral, I think itâs high time alternative strategies are explored and new approaches tested.
Have a great evening,
D.
đ¶ listening to the âMachine I-Vâ box set from The Bug. Newly signed to Relapse Records, this is quite the packaging gem, rolling up the artistâs five âMachineâ releases into a set of 12âs. Wonderful stuff.
đș watching Amoebaâs latest âWhatâs In My Bagâ clip with Marc Maron. Maron could be a writer for The Quietus, such is his range of tastes that takes in anything from Can to My Bloody Valentine to Boygenius. Well worth a watch.
đ€ playing with the podcast creation abilities of Googleâs Notebook LM. It is great, if a little unsettling at points. Donât believe me? Hereâs itâs podcast of my writing above. What cannot be denied is how impressive this is though:
Stories worth reading from the Music Industry:
Inside TikTok's Money Squeeze on Independent Labels
âWe have worked productively and collaboratively with TikTok on this [fraud] issue, and until now, no concerns have been raised about the approach Merlin is taking. We have implemented measures to address illegitimate activity and content; automated systems to detect suspicious activity; and a dedicated team to address issues and impose sanctions, including ultimately, termination of membership for bad actors.â These claims were seconded by other sources, who also note that music is uploaded directly from labels onto TikTok, without passing through Merlin or any middleman.
đđ»Hot take: this feels like a blatant divide-and-conquer tactic from TikTok, I do wonder if the ease with which Merlin has been cut out, along with the potential impact of that, might suggest that this is another example where trade bodies simply arenât powerful enough now to fully represent rights holders.
Music creator economy: recalibration
Streaming has been through a number of cycles with creators: 1) distrust; 2) enthusiasm; 3) royalty dissatisfaction; and now 4) apathy. Our latest music creator report, which presents data from MIDiAâs latest global music creator survey, shows that progressively more creators are starting with lower expectations for streaming. Their royalty expectations are already so low that this is no longer a pain point for them. Instead, they are becoming critical of streamingâs ability to further their careers, focussing on the mediumâs closed door between them and their fans. Todayâs creators want platforms where they can build high value, smaller fanbases, rather than low value, large-but-anonymous audiences.
đđ»Hot take: these findings entirely reflect what I am seeing, and what Iâve been writing about for some time now. DIY-tier artists just donât care about streaming any more. It does nothing for them.
GEMA proposes licensing model for AI-generated music
Through its new licensing model, GEMA proposes that authors should receive compensation beyond a one-time payment for training data. It suggests that such one-off payments may not sufficiently compensate authors given the potential revenues from AI-generated content. GEMA says its model advocates for âfair remuneration at a high level while keeping in mind that the market and its technical developments can change dramatically and rapidly.â
đđ»Hot take: this is commendable, but entirely in opposition to what AI companies are looking for. Landmark cases are needed in this space to set precedents and (ideally) force the hand of Big Tech.
Ex-Island boss Darcus Beese: âThe major-label ecosystem is brokenâ
âTodayâs music is tomorrowâs catalogue. And I donât want to be part of the most anaemic part of history. When archeologists come around and go âwhat happened in this time period?â⊠and you go âWhat happened here? Yeah, TikTok. People didnât give a fuck. â Todayâs music is tomorrowâs catalogue, and rec catalogue should live 10 and 20, 30, 40, 50, 60 years. I donât know in this last five, eight years whether weâve all been doing that⊠Thereâs not enough inspiration and creativity around at the moment.
đđ»Hot take: whilst something sticks in my throat about execs who make a living from major labels, only to then leave and take a highly critical stance against their former employer, there are nonetheless some really interesting points made by Beese here. Iâd certainly agree with his comments about A&R, artist development and TikTok.
Notable news from the world of tech:
What feels real enough to share
The Vergeâs Nilay Patel recently summed up the core tension here, writing on Threads about YouTubeâs own generative-AI efforts, âEvery platform company is about to be at war with itself as the algorithmic recommendation AI team tries to fight off the content made by the generative AI team.â And itâs clear, at least with Meta, which side is winning the war. This week, Meta proudly announced a new video-generating tool that will make AI misinfo even more convincing â or, at least, better at generating things that feel true.
đđ»Hot take: behold, the era of AI slop. As the pull quote suggests, we are now moving into a faintly ludicrous period where at one end you have AI-generated dross flooding social media, and at the other you have the social media companies trying to pretend this isnât a serious issue. It is absurd, and it concerns me the degree to which everyone is going along with it.
Companies building AI-powered tech are using your posts. Hereâs how to opt out
To make it easier for you, weâve gone through the painstaking process of figuring out how to pull your data out of the hungry AI machine on a bunch of the major platforms so you donât have to wade through endless settings and blocks of text to figure it out yourself. While some opt-out processes were straightforward once we knew what we were looking for, others were much more convoluted. By far, Meta was the most difficult platform on which to find ways to protect our data. And if youâre in the US, there is currently no way to stop the company from training its AI on your information.
đđ»Hot take: I included this mainly as something of a PSA to readers. As with most things around AI it feels like major legal judgements are still to be made as to the degree to which our content is being used to train LLMs and generate more money for yet another big tech company. Where does this end?!
Google ordered to open Play store to rivals after antitrust loss to Epic Games
A US judge on Monday ordered Alphabetâs Google to overhaul its mobile app business to give Android users more options to download apps and to pay for transactions within them, following a jury verdict last year for the Fortnite game maker Epic Games. The injunction by US district judge James Donato in San Francisco outlined the changes Google must undertake to open up its lucrative app store, Play, to greater competition, including making Android apps available from rival sources.
đđ»Hot take: I will be interested to see if this compels someone like Epic to provide and app store that might, for example, provide a workaround path to devices for the likes of Bandcamp. Certainly this could all have ramifications for music services and what weâre paying for them.
Looking for something else to read? Here you go:
Drowning in Slop
A thriving underground economy is clogging the internet with AI garbage â and itâs only going to get worse.
đđ»Hot take: another great take on the AI slop era, eminently worth your time.
Forums Are Still Alive, Active, And A Treasure Trove Of Information
Here are the Internet forums that are still alive and kicking and full of information and interesting people.
đđ»Hot take: I loved this - just a gentle reminder that whilst some obsess over social media, the world of forums is still very much alive.