It’s over for human-generated content.
That seems to be the conclusion of Meta’s AI consultant, Henry Ajder, who recently warned that the internet will soon be taken over by “AI slop” — insta-generated, low-quality, noisy, and derivative text, video, images, and even influencers.
All to make numbers go up while also crushing signal fidelity, creativity, and information quality.
Meanwhile, Meta corporation remains locked in its own historical pattern of scrambling to retain relevance in the digital cash-grab — now manifested as the race to AI market dominance. This race presents a circular conundrum: The more output from AI tools you create, the more AI tools you need to manage it.
If Ajder is correct, then using the internet is about to get even more tiresome. For how long, no one, not even Ajder, knows.
‘The age of slop is inevitable. I’m not sure what to do about it.’
The reasoning is all shortsighted, but not at all mysterious. It’s akin to the constant call for weapons upgrades in a military context: If they get it before us, we’re dead.
It may be the case that this base impulse to compete in austere, non-human market conditions has much to do with the rather shameful and sordid historical arc of Meta and its leader, Mark Zuckerberg. A pattern emerges where the humanity at play seems to be something along the lines of a child who feels left out and wants to play, and shorn of morality or bearing, he will do anything to get the chance to be cool.
A decade of scrambling
Consider the arc: Meta really begins with Facebook and leaps up with the acquisition of Instagram. There’s no argument in retrospect to overcome the all-but-catastrophic effects, the sweeping social decay, and human suffering by dispersal and rewiring of human attention spans that are the true fruit of this social media empire.
With all the data, illegally and immorally harvested perhaps, a switch into AI for Zuckerberg was in the cards from day one. By the time Instagram was purchased in 2012, the big push was toward machine learning. The same year saw Google drop big money into AI via Google Brain.
Between 2014 and 2018, the business and development ecosystems surrounding artificial intelligence take shape. OpenAI emerges, famously. The General Data Protection Regulation emerges in European law as a protection against privacy and data-rights violations. In this period, Zuckerberg/Facebook acquires messaging platform WhatsApp and VR company Oculus while also marking up big numbers in terms of violating data and privacy rights.
What does Zuckerberg do? He adds AI content moderation. It’s clear there’s neither a moral center nor some visionary, religious-level plan. It’s just grabbing the bag, over and over again. Chasing big brother, as it were.
In 2021, the virtual Zuckerberg-a-verse rebrands as Meta. The fact is that Facebook had become a sort of pre-slop disaster scene of ads, scams, and bots. A mass exodus from the platform by Gen X and younger users is memorable to all who lived through the era. The rebrand to Meta signals a shift, perhaps formally, into AI, but it’s late and it feels without focus or design.
RELATED: Zuckerberg to dump hundreds of billions into new Manhattan-size projects
Craig T. Fruchtman/Getty Images
Llama drama
Perhaps it was the case that Zuckerberg just needed his peers to point the path. By 2022, Meta developed an open-source AI model called Llama, but it wasn’t until 2023, amid that project’s underperformance, that the company began to pivot in earnest.
This year, Meta has gone on a conspicuous hiring spree to buy its way to the top of the AI game. The company has hired veterans from OpenAI, DeepMind, and Anthropic and turned the main project, the holy grail, apparently, that all AI companies are crawling toward, over to Alexander Wang. Wang is tasked with bringing the Meta AI capacities beyond human levels, into the realm of so-called true superior intelligence.
Meta is still dragging Facebook, Instagram, WhatsApp, and a suite of quasi-metaverse/virtual reality products to market with, let’s say, varying degrees of efficacy and social or historical importance. Consistent with the pattern of simply keeping up with the herd, avoiding the bold and fearsome mess of human originality, Meta’s stab at the ultimate “reasoning model” LLM is simply called Superintelligence Lab.
If Meta’s conjuring up of ultimate intelligence pans out, it will be a story of coming from behind. And given Henry Ajder’s slop warning, it may be that Zuckerberg will need to, once again, turn quickly toward emergency triage and apply his AI to a cyber disaster zone effected, in large part, by other AI models.
Flood warning
This time, not just Facebook, but great swaths of the internet as a whole face a cataclysm. Here’s what we’re facing when and if the flash flood of AI-driven super-slop arrives.
- Category 5 slop storm hits as predicted in late 2025.
- 50% of all social media content is AI-generated or automated.
- Use cases dominated by low-quality text, images, videos, and ads.
- Trust crashes as no standards develop for authenticating human vs. AI content.
- Meta’s own tools are turned on swamping its platforms with slop.
- Markets grow for generative models fitted to evade AI security detection.
- Users and companies scramble to find cryptocurrency solutions to authentication.
But in spite of the dawning recognition that Bitcoin offers a scalable way to distinguish human from automated artifacts, Ajder states “the age of slop is inevitable.” He sighs, “I’m not sure what to do about it.”
So we’re probably in for yet another ripple in the story of the evolving internet, but perhaps there’s a bright light, maybe not for Meta but for the rest of us normal people, and it’s this: If the big platforms are shattered by their own serpentine machinations gone awry, a market may open for smaller, curated, niche platforms where genuine human online interaction can make another play.
Being that there is no real discussion to either moderate slop or to break the monopolistic big tech entities, this author would, all things considered, be willing to wade through an era of turbo-tedium for a chance at platforms more human, with better boundaries, and perhaps built by founders interested more in human interaction and less in simply chasing trends, number lines, and grand delusion.
Read the full article here
