01
AI slop detection does not exist.
Token math, variance, and burstiness can estimate whether writing is predictable. They cannot reliably tell whether a sentence is linguistically likely to be Unnaturally Recurring Direct AI Output, or AI slop.
why antislop
AI is everywhere. your voice should be too.
Humans are already using AI to automate writing at unprecedented scale. The experience of the internet is changing with it: behind every expression used to be an individual, with their unique takes, quirks, and personality. We built Antislop because voice and written text are, and always will be, important byproducts of humanity.
who we are
slop is a choice now.
“slop is the greatest threat to humanity's speech since social media.”
We should, must, evaluate who we are, what our purpose is, and how we can collectively shape AI toward the benefit of humanity. Our ideas are small, yet that gives us hope and power, because all great movements start small.
Akasha Montague
founder, adeimantos
the priority
AI text is not the whole problem. voice is.
Every platform is being overrun with text. LLMs produce text every day, and individuals, businesses, farms, and mass bot setups are publishing AI-written content to generate views and revenue. It is more a human problem than an AI problem: the choice is to use text as it is generated, or spend time structurally rewriting entire pieces when speed is the moat and revenue is the bottom line.
For us, AI text is inherently not the problem. It is how the text was edited or regenerated before it was posted. The problem, and the priority for us, is voice. We are losing the expression, personality, and individual perspective that made the internet feel human in the first place.
how others approach this
three familiar promises. one missing layer.
02
AI humanisers do not exist.
A system that rewrites the patterns it happens to flag makes some text less predictable to a detector. It does not isolate the linguistic spans that need to change while keeping the basis, points, and facts of the text intact.
03
“Sound like you” AI does not exist.
Adding a voice description to a whole-text rewrite does not change the model’s behaviour. The beginning may sound like you; by the end, the context has deteriorated and the output carries the same smooth-brain effect.
what we do
find the slop. keep the person.
Antislop's in-house detector is purpose-built for structural linguistic pattern detection with a dual-layer verification system that catches real slop and lets non-offenders through. We care whether every sentence was written to sound human, rather than whether a machine can identify the overall text as AI.
Antislop balances two sides of the same coin: millions of AI users can produce copious amounts of content, drowning out the voices of hundreds of millions of real humans. We reconstruct what needs to change while preserving the source and the voice behind it.
the future
more capability. more humanity.
Antislop is positive about the use of AI in our society: individuals from around the world can bring their ideas into reality where they were once gated behind long degrees and learning curves. We can only hope, we must hope, that the individual voice never gets drowned out. That is why we are here, and why you are too.