Free tools / AI Slop Phrases

The list our own writing is checked against

AI Slop Phrases

These are the phrases that make copy sound machine-written. It's the live list behind our AI Slop Checker and the gate our own blog, our customers' drafts and our product screens pass before they ship. When we hear a new one, it goes on the list, and this page shows it the same day.

102 phrases in 12 groups, read live from the list that gates our own writing.

Withheld tease

Promising a payoff instead of delivering it: what they don't tell you.

  • what they don't tell you
  • what they didn't tell you
  • what nobody tells you
  • what they did not know
  • what came next
  • what happened next
  • what you'll hear next
  • what you will hear next
  • blew everyone's mind
  • blew everybody's mind

Pause for effect

Let that sink in: a stage direction where a point should be.

  • let's sit with that
  • let us sit with that
  • let that sink in
  • sit with that for a moment
  • for a moment
  • for a minute

"Worth" constructions

Worth noting, worth knowing: a hedge dressed as emphasis.

  • worth knowing
  • worth noting
  • worth doing properly
  • worth reading closely
  • worth saying
  • what's worth
  • it's worth

"That decides"

That's what decides: a drumroll before the answer.

  • that's what decides
  • that is what decides
  • which is what decides
  • because that decides

"Nobody told them"

The underdog arc: nobody told them, they did it anyway.

  • nobody told them
  • nobody said it
  • no one told them
  • they did it anyway

Imperative openers

Picture a buyer, start with the thing: an order to imagine.

  • start with the thing
  • start with what
  • now picture
  • picture a buyer

False humility

To be honest, full disclosure: implies the rest was not.

  • i'll be honest
  • to be honest
  • i have to admit
  • full disclosure

Claude vocabulary habits

Words Claude models are known to reach for: tapestry, testament to, delve into.

  • tapestry
  • testament to
  • delve into
  • delicate balance
  • a testament
  • rich tapestry

ChatGPT vocabulary habits

Words ChatGPT models are known to reach for: furthermore, crucial role, in summary.

  • furthermore
  • in summary
  • crucial role
  • in today's fast-paced
  • robust and scalable
  • it's important to note

Gemini vocabulary habits

Phrases Gemini models are known to reach for: here's the deal, let's break it down.

  • here's the deal
  • think of it this way
  • let's break it down
  • let's dive into
  • the bottom line

Inflated adjectives

Pivotal, transformative, game-changing: size words with no measurement.

  • pivotal
  • game-changing
  • transformative
  • cutting edge
  • seamless integration

Stock phrases

Throat-clearing and filler that every style guide bans: moreover, game-changer, in today's.

  • in today's
  • in the world of
  • in an era
  • it's important to note
  • it is important to note
  • it's worth noting
  • needless to say
  • at the end of the day
  • when it comes to
  • the fact of the matter
  • let's dive in
  • let's explore
  • buckle up
  • game-changer
  • game changer
  • revolutionize
  • revolutionise
  • unlock the power
  • harness the power
  • leverage the
  • in conclusion
  • to sum up
  • the bottom line is
  • moreover
  • furthermore
  • additionally,
  • delve into
  • navigate the
  • landscape of
  • ever-evolving
  • fast-paced
  • cutting-edge
  • seamless
  • robust solution
  • holistic
  • synergy
  • paradigm shift
  • not only
  • isn't just
  • is not just
  • more than just

How it works

How the list works

  1. One phrase per line, plain English

    No regex. Anyone on our team adds a line when they hear a phrase that gives copy away, and the checker compiles it into a whole-word match.

  2. Grouped by the problem, not the phrase

    A category says what kind of failure it is, so a writer fixing a hit knows why it reads badly.

  3. One test for adding a phrase

    Would you say it out loud to a customer, in a meeting? If not, it goes on the list.

  4. The checker runs more than this list

    Stock phrases, jargon and meta-commentary rules are built into the scorer, along with five sentence shapes no phrase list can express.

Reference

The five sentence shapes no phrase list can catch

Structural patterns the AI Slop Checker detects in addition to the phrases below
ShapeLooks likeWrite instead
The 'X, not Y' flipUse a comma, not a dash.Use a comma.
Tail clauseWe measured it, and that is the point.We measured it.
This-is openerThis is why it matters.Name the thing: Slow replies cost billable hours.
Possessive tailA formula, and it is mine.My formula.
Em dashThe page loads, then stalls with a dash between.Two sentences, or a comma.

Why publish the list?

Because a list you can read beats a score you can't argue with. Every flag our checker raises traces to a line here or a rule we name on the checker's page. If a phrase is on the list and you think it shouldn't be, you can see exactly what you're disagreeing with.

It's also how we keep ourselves honest. Our own drafts fail on these phrases often enough that the list keeps growing. We'd rather publish the standard and meet it in public than claim a voice we don't check.

Will removing these phrases make AI cite my page?

No checker can promise that, and we won't. Removing them makes copy easier to read and easier to trust. Engines choose pages for what they say, whether they answer the query a buyer typed, and who else backs them up. Use the list to clean the writing, then measure whether engines name you.

Clean copy is the floor. It can't tell you whether AI names you.

A free scan puts real buying queries to all five AI engines and shows who they name instead of you.

Check your AI visibility free →

FAQ

Frequently asked questions

What is AI slop?

Copy that reads generated: filler, stock phrases, teases that withhold the point, and the same few words every model reaches for. It's a description of how text reads, whoever wrote it.

Why does 'worth noting' count as slop?

It's a hedge dressed as emphasis. If a point is worth noting, note it. The 'worth' family runs through generated copy and through ours, which is why it's on the list.

Are the model vocabulary lists proof a model wrote something?

No. They're words each family is known to overuse. People use them too, and a model can be told to avoid them. The checker counts them as habits and never names an author.

Why isn't 'delve' the only word here?

One famous word is easy to avoid. The tells that survive editing are structural: the tease, the pause for effect, the 'not X, but Y' frame. Those are most of this list.

How often does the list change?

Whenever we hear a new phrase. This page reads the file on every view, so the version here is the one the checker runs today.

Can I use this list for my own team?

Yes. Link to this page, or run your drafts through the AI Slop Checker, which marks each phrase in place and names its category.

Does the checker flag every word on this list everywhere?

It matches whole words and phrases, so 'tapestry' is flagged and 'tapestries' isn't. Straight and curly apostrophes both match.

Is this list free to use?

Yes. The checker, the list and every other deterministic tool on this site are free with no signup. What we sell is measuring whether five AI engines name you when buyers ask.

Sources

Sources and further reading

  • stop-slop: the published ruleset our checker started from