We have a new insult: AI slop.

You’ve probably seen it. A suspiciously glossy image appears online and somebody calls it slop. An AI-generated song shows up in a feed. Slop. Someone discovers that generative AI helped write an article. Slop. A video, advertisement, illustration, website, poem, presentation, story, photograph, voiceover or meme has generative fingerprints on it, and sometimes the verdict arrives before anyone has even considered the thing itself.

AI made it. Therefore, slop.

There is certainly plenty of AI slop in the world. Generative systems have dramatically reduced the cost of producing content, which means they have also dramatically reduced the cost of producing enormous quantities of mediocre content. Search-engine filler, engagement bait, disposable social posts, endless variations of essentially the same image, and articles nobody needed written primarily so another webpage can exist. The slop factory is real, and generative AI has given it a very efficient assembly line.

But there is a problem with treating “AI-generated” and “AI slop” as synonyms.

Humans make slop too.

We always have. Bad novels existed before large language models. Terrible music existed before generative audio. Forgettable advertising existed before image generators. Hollywood has occasionally managed to spend hundreds of millions of dollars producing something audiences forget before reaching the parking lot. Human authorship has never guaranteed quality, originality or emotional impact.

So perhaps the arrival of generative AI requires a different test. Instead of beginning with Who made it? or even How was it made?, maybe we should begin somewhere much closer to the actual experience of art and media:

Does it resonate?

That question moves the conversation somewhere much more interesting, because creative work has always involved more than its creator. There is whoever or whatever creates the artifact, and there is the person who encounters it. We spend enormous amounts of time talking about the first. Artificial intelligence may force us to think much harder about the second.

The Other Half of Creativity

Imagine hearing a song without knowing anything about its origin. There is no album cover, artist biography, record label or story about someone writing it at three in the morning after a breakup. You simply hear the music.

Something happens.

Maybe the melody reminds you of someone. Maybe one line lands harder than you expected. Maybe you replay it. Maybe you send it to a friend. Maybe, for reasons you cannot entirely explain, it becomes attached to a particular period of your life.

Then someone tells you the song was generated by AI.

What happens to the experience you already had? Does the emotion disappear retroactively? Was the memory counterfeit? Did your nervous system somehow need an authenticity certificate before deciding whether the song affected you?

Whatever happened between you and the song has already happened.

That does not settle every question surrounding AI-generated creative work. Copyright, compensation, training data, attribution, labor displacement and cultural homogenization are serious issues deserving their own arguments. But those are different questions from the one we are asking here.

The question here is simpler: Did the thing reach you?

Meaning does not exist exclusively inside the creator. Meaning also happens inside the audience.

The Creator May Not Control the Meaning Anyway

This isn't a problem invented by artificial intelligence. Humans have been arguing about artistic intention for a very long time. Writers discover interpretations of their novels they never consciously intended. Listeners attach deeply personal meanings to songs that have little relationship to what the songwriter originally had in mind. Paintings become symbols of movements their painters never anticipated.

Once a creative work enters the world, the creator loses exclusive control over what it means. The audience begins participating in the creation of meaning.

Every person brings memories, experiences, culture, fears, desires and associations to the encounter. Two people can hear precisely the same song and experience two entirely different things. One might hear a catchy melody. Another might hear something that instantly transports them thirty years backward.

That creates an uncomfortable possibility when we introduce generative AI:

Meaning may not require meaning on the creator's side.

Suppose an autonomous generative system produces a song. No human writes the lyrics, composes the melody, chooses the instrumentation or carefully selects among twenty generations. For the sake of the thought experiment, remove meaningful human creative intervention almost entirely.

The machine produces the song.

Then a human hears it and cries.

Where is the meaning?

Perhaps there was none inside the machine. The machine did not experience heartbreak. It did not remember someone it lost. It did not lie awake thinking about the words it had written.

But something unquestionably happened inside the human listening to it. The machine does not need to experience heartbreak for a particular arrangement of words and music to remind you of yours.

The creator may have felt nothing. The audience did.

That distinction is going to become increasingly difficult to ignore.

But Most AI Creativity Isn't Actually That Simple

There is another problem with the phrase “AI-generated.” It describes an enormous range of creative processes as though they were the same thing.

They aren't.

Consider someone who types, “Make a country song about Wisconsin,” accepts the first result and uploads it. Now compare that with someone who spends hours working with generative systems. During that process they might:

• Research the subject and develop a narrative.
• Decide which ideas, events or emotions matter.
• Experiment with genres, voices and structures.
• Reject drafts and rewrite sections.
• Correct factual errors and pronunciation problems.
• Rearrange the sequence and emotional arc.
• Generate alternatives, listen to them, react and revise.
• Keep iterating until something finally triggers that peculiar internal response: Yes. That's it.

Calling both processes simply “AI-generated” eliminates almost everything interesting about how the second work came into existence.

Increasingly, creative production with generative systems is not a straight line from human prompt to machine output. It is a loop: Human ? AI ? Human ? AI ? Human ? Artifact.

The human changes the machine's output. The machine changes the human's thinking. The human responds to something unexpected. The next generation incorporates that response. A direction appears that neither side contained at the beginning, and the eventual artifact emerges from the interaction.

At Conjugo, we have been calling this a human-AI dyad.

Less formally: MeatBot + MagicBot.

The terminology isn't particularly important. The process is, because something unusual happens when creation becomes conversational.

Creation Becomes Recursive

Traditional tools generally wait for us. A guitar does not suggest a chorus because it thinks the second verse needs one. Photoshop does not normally disagree with your concept. A word processor does not read your paragraph and propose an entirely different thesis.

Generative systems can.

That makes them strange creative instruments. You can begin with an idea and receive something unexpected. That unexpected output changes your idea, so you respond. The system responds to your response. The original concept mutates.

After enough iterations, it can become genuinely difficult to identify exactly where particular ideas originated. You might find yourself wondering:

• Was that line originally mine or the AI's?
• Did I suggest the concept and the AI transform it into something else?
• Did the AI misunderstand something in an interesting way that sent the project in a better direction?
• Did something emerge through the exchange that neither participant could reasonably claim as the original idea?

Eventually, those questions may become less useful than they initially appear. The provenance becomes tangled, but the artifact remains.

Then someone else encounters it.

Which brings us to the only person who wasn't involved in the creation process: the audience.

The Audience Doesn't Owe the Creator Anything

Here is where the argument cuts both ways. Human-AI collaboration does not deserve special treatment simply because the process was elaborate. Nobody has to appreciate something because a human spent twelve hours prompting it. Nobody has to admire an AI-generated song because its production workflow was sophisticated. Nobody has to care that forty-seven drafts were rejected.

Effort is not a quality guarantee. Neither is technology.

The audience owes us nothing.

They encounter the work and decide whether to keep going. Their internal evaluation might be remarkably simple:

• Did this make me think?
• Did it make me laugh?
• Did I learn something?
• Did it surprise me?
• Did I feel something?
• Do I want another thirty seconds of it?
• Would I send it to somebody?
• Will I remember it tomorrow?

In other words: Does it resonate?

If the answer is no, perhaps it really is slop. Or perhaps it simply isn't for that person. Human culture has always worked that way.

The Provenance Paradox

There is a revealing experiment hiding inside all of this.

Show someone an artwork without identifying its origin. Let them experience it first. Only afterward reveal how it was created: human, AI, or human-AI collaboration.

Would their evaluation change?

For some people, almost certainly.

And that tells us something important. We do not judge creative artifacts solely by experiencing them. We also judge the stories surrounding them.

A handwritten letter feels different because someone physically wrote it. A live musical performance carries information beyond the sound itself. Knowing that a painter spent years developing a technique can affect how we experience the painting. Knowing that a songwriter wrote something after losing someone can fundamentally change the emotional context of a song.

Provenance matters. Human effort matters. Intent matters. History matters.

But they are dimensions of the experience, not necessarily prerequisites for having one.

An AI-generated song and a song written by someone mourning the death of a parent may sound equally beautiful while carrying radically different human contexts. We don't need to erase that distinction. We simply don't need to pretend that one prevents the other from affecting us.

The Coming Flood

This question becomes more important because generative content is becoming extraordinarily abundant. Music, video, writing, games, advertising, virtual worlds, interactive characters and eventually entire media experiences can increasingly be generated cheaply and rapidly. Some of those experiences may eventually be created specifically for an audience of one.

The scarcity that defined much of human media production is weakening. Soon the problem may not be creating things at all. The problem will be finding things worth experiencing.

In a world containing effectively unlimited content, provenance alone becomes a poor filtering mechanism. We will need other signals:

• Trust and reputation.
• Taste and originality.
• Human or algorithmic curation.
• Community and shared experience.
• Context and cultural relevance.
• And perhaps the oldest signal of all: Did this thing move me?

Ironically, the explosion of machine-generated media may make human judgment more important rather than less.

Machines can generate the ocean.

Humans still decide what is worth pulling from it.

Maybe “Slop” Isn't About AI at All

Perhaps we have misidentified the enemy.

Maybe slop isn't content created by artificial intelligence. Maybe slop is content created without sufficient regard for the person receiving it.

It is content designed merely to occupy space, manipulate an algorithm, harvest attention, imitate whatever performed well yesterday, or generate another impression, click or pageview. It exists because somebody wanted another unit of content rather than because somebody believed there was something worth communicating.

Machines can certainly produce that.

So can corporations, marketing departments, media companies and individual humans. We have centuries of evidence.

The technology changes. The underlying failure doesn't.

Nobody cared whether the audience received something meaningful. They only cared that something was produced.

That may be a much better definition of slop.

The Test

So here is a proposal.

The next time you encounter something made with generative AI, temporarily forget how it was made. Look at it. Read it. Listen to it. Watch it. Experience it.

Then ask yourself:

• Did it make me feel anything?
• Did it make me think?
• Did it show me something differently?
• Did I enjoy the experience?
• Will any part of it remain with me?

If the answer is no, discard it. The world is not suffering from a shortage of content.

But if the answer is yes, something more interesting has happened. A collection of tokens, pixels, frequencies or frames crossed the boundary between artifact and mind. Something resonated.

Maybe a human created it. Maybe an AI created it. Maybe, increasingly, neither description adequately explains what happened. Maybe a human and an artificial intelligence spent an afternoon throwing ideas back and forth until neither could cleanly identify who contributed what.

We are going to spend years arguing about what to call that: art, content, synthetic media, AI-generated media, co-created work, slop or something we haven't invented a word for yet.

Those labels matter.

But they aren't the first question.

The first question belongs to the person on the other side of the work: the listener, the reader, the viewer, the human.

And their question is beautifully primitive:

Does it resonate?

If it does, maybe that's where the interesting conversation begins.