AI Slop is Not New, Just Industrialised

The behaviour came first, the label arrived later.

For many years, lazy and shady people have been building “channels” by scraping other people’s clips, Reddit posts, pet videos, podcast moments and news footage, then stitching them into compilations with a new title, thumbnail and maybe a watermark.

AI simply took that lazy mentality and put it on steroids.

What “AI slop” actually means

“AI slop” is now widely used to describe cheap, low-quality AI-generated text, images and video pumped out in huge volumes with minimal human input.

Analysts and journalists describe this trash pile-up as an evolution of spam – a mix of engagement bait (often ‘rage bait‘), misinformation and filler that crowds out genuine human work.

Researchers estimate that over half of longer English posts on platforms like LinkedIn are now AI-written, much of it formulaic boilerplate.

So when you complain about slop, you’re not overreacting. You’re giving a name to something people across tech, media and academia are already worried about.

Lazy “public service” and logo-slap channels

The same low-effort mentality shows up in places that like to present themselves as serious or reputable. Take live coverage of parliamentary debates and PMQs:

The official Parliament channel already streams and archives PMQs and key debates as part of its remit. Big news outlets stream the same events – slapping their own logo in the corner, of course, and embed them in studio programmes with presenters, commentary and analysis. Sometimes with sod all extra added.

Then you get a long tail of “politics” channels that simply re-stream the official feed, put their own logo in the corner, and upload the whole thing as if it’s their own work. YouTube watch-time revenue makes people do whatever they can to ride any bandwagon.

The awkward bit the defenders don’t like.

Public footage doesn’t mean you magically become a creator just because you captured the same feed and added a sodding logo. Copyright and fair use principles explicitly hinge on transformation and purpose: commentary, criticism, context, or new use – not bare duplication.

Best-practice guidance for online video is blunt about it: fair use is strongest where you add meaning, weakest where you just mirror.

So when someone claims re-uploading PMQs with their logo is a “public service”, I call bullshit. They’re monetising other people’s work while pretending it’s a civic duty.

Why so much slop exists: the incentive problem

The uncomfortable truth is that slop is rational behaviour in a warped system.

Platform algorithms reward content that keeps people scrolling and tapping, not content that’s well-researched or useful.

Studies and reporting show that Facebook in particular is promoting viral AI-generated images and posts (including surreal things like “shrimp Jesus” – yes, honest) instead of suppressing them, because they perform well as engagement bait.

For a few quid, anyone can buy thousands of fake likes, views and AI-generated comments, making slop look popular and pushing it further up recommendation systems.

If you’re an opportunist with no real expertise or craft – and can’t be arsed doing any thinking or working, the message from the system is simple:

• shove out as much cheap, effortless content as possible,
• remix or scrape whatever’s trending – sod the copyright,
• bolt on an AI model to write scripts, comments or fake chat,
• and funnel the resulting traffic into courses, memberships or affiliate links.

I value graft and responsibility and here’s the core problem: platforms and tools have made it easier than ever to build an income stream without doing any meaningful work or taking any real responsibility for what you publish.

“It has always been rubbish” and other counterpoints

To stay honest, some pushback is worth acknowledging.

“The web was full of crap before AI”

A fair criticism and fairly true.

Plenty of commentaries and forum posts point out that we’ve been swamped with shallow clickbait, low-effort product reviews, and content-farm articles for years, long before “AI slop” became a phrase. In that view, AI hasn’t broken a golden age – it’s just made an already messy internet more obviously broken.

The important distinction is scale and recursion.

Human slop is annoying. AI slop can be produced in industrial quantities, and then (it gets worse) that same slop gets used as training data for the next generation of models. Researchers warn this feedback loop leads to “model collapse“, where systems trained on their own output become less accurate, more biased and more nonsensical over time. That’s a different order of damage.

“Most people only want quick, snackable stuff”

There’s truth here too.

People do consume huge amounts of short, disposable content. Platforms argue they’re responding to consumer demand: short clips, bold thumbnails, surface-level takes. In that sense, slop is partly a mirror of our own habits.

But convenience isn’t the same as consent.

Most normal users don’t ask to be fed fake AI “news” footage, synthetic influencers or invented Reddit drama that’s indistinguishable from real human conflict. They want quick content, not carefully crafted lies.

“AI can help good creators, not just grifters”

Also true.

Used properly, AI can handle subtitling, first-pass drafts, language support, basic editing and accessibility, especially for small teams and solo operators. Some journalists and creators are using it to expand coverage, not replace human judgement.

The line is intent and effort.

If you start with expertise and a point of view, AI can make you faster.

If you start with nothing, AI lets you mass-produce nothing in a more convincing shape.

My critique is aimed at the second camp, not the first.

The “it will fix itself” optimism

There’s a hopeful camp that says this is just another phase like email spam.

Some experts argue that as people get bored and more tools arrive to detect AI content, the worst slop will become less visible.

Platforms promise “robust defences” against low-quality or near-duplicate posts, and some already downrank obvious spam.

Maybe. But look around.

The volume of AI-generated output is rising far faster than the quality of filtering, and there’s serious money to be made in pumping it out. For now, the economic incentive to keep shovelling slop massively outweighs the incentive to clean it up.

From a balanced standpoint, the fair position is:

• yes, filtering can improve,
• yes, some of this will self-correct as users push back,
• but no, there’s nothing inevitable about that if regulators, platforms and audiences all look the other way.

The real cost: trust, not just aesthetics

This is bigger than “I’m annoyed by low-quality posts”.

When newsrooms and high-visibility brands quietly deploy AI for articles or bylines without clear labelling, they chip away at trust in every headline and byline, not just the bad ones.

When social feeds fill with convincing fake videos of real people doing things they never did, the default response becomes scepticism of everything, including real evidence and genuine journalism.

As more high-quality work retreats behind paywalls to survive, the free public internet risks becoming a slurry of AI slop for the masses, with serious information reserved for paying customers.

That’s where democracy, accountability and basic civic life get dragged into it.

If most people’s daily information diet is sourced from an endless scroll of cheap, unaccountable AI content and lazy reuploads of “public” footage, then their grip on reality weakens. Not because they’re stupid, but because the signal is buried in noise.

Where this old bugger’s opinion lands

If you come at this valuing personal responsibility, graft, and honest dealing, AI slop offends on several fronts at once:

• It normalises getting paid without real work or expertise, by dressing up scraping and remixing as “content creation” or “public service”.

• It lets big platforms and businesses cut costs by flooding the world with cheap output, while offloading the long-term cost of confusion, fraud and distrust onto everyone else.

• It erodes the basic ability of an ordinary person to tell whether what they’re seeing is real, human, and accountable.

Balance doesn’t mean pretending this is fine.

Balance means admitting that tools can be used well, that some creators and journalists are genuinely trying to use AI responsibly, and that human junk existed long before AI. Then it means saying, plainly, that what we’re seeing now is something worse: a structural shift towards an internet where slop is the default and craft is the exception.

The internet doesn’t have to be a trough.

But if we keep rewarding people for shovelling muck into it, we shouldn’t be surprised when decent, human work gets trampled, and ordinary users decide they’ve had enough of the whole bloody thing.

A Few FAQs about AI Slop