Sentence Openers Audit: Why Starting Every Line With ‘The’ or ‘This’ Looks Generated

Read a paragraph of AI writing out loud and you’ll often hear a tic before you can name it: every sentence seems to lurch off the line with the same word. The. This. It. The. This. It’s the prose equivalent of a marching band, and once you notice it you can’t unhear it. Worse, it’s one of the cleanest signals that a draft came out of a machine.

So let’s run an audit. Not on your grammar, not on your vocabulary — just on the first word or two of every sentence. It’s a five-minute pass that does more for how human your writing sounds than almost anything else you can do quickly.

Key takeaways

• AI drafts lean hard on a few openers — “The,” “This,” “It,” and the participle phrase — and the repetition is one of the easiest tells for a reader to feel and a detector to measure.
• The fix isn’t banning a word; it’s variety. Count your first words, and if one shows up at the top of three sentences in a row, change one of them.
• Strong alternatives include leading with the subject’s actual name, a short time or adverbial phrase, a question, a dependent clause, or a blunt short sentence.
• “This” with no noun after it (“This shows…”) is doubly suspect: it repeats an opener and leaves the reference vague. Add the noun (“This gap shows…”).
• Vary openers to serve the reader, not to dodge a tool — detectors report probabilities, not proof, so treat the audit as a craft habit.

Why repeated openers read as generated

Language models choose words by probability. At the start of a sentence, the highest-probability, lowest-risk opening words are a small club: *The, This, It, In, There,* plus the “-ing” participle phrase (“Considering the data…”). Nothing in the model’s default behavior pushes it toward variety, so it keeps reaching for the same safe starters. The result is a paragraph where the openings barely change.

Human writing doesn’t do that, at least not for long. We start with a name, then a question, then a fragment, then a long winding clause, because we’re following meaning and emotion rather than picking the statistically safest next token. That unevenness is part of what makes prose feel alive — and it’s part of the statistics detectors actually measure when they score the variation in your sentences.

One honest caveat before we go further: detectors output a probability, not a confession. OpenAI even retired its own AI-text classifier in July 2023, citing a low rate of accuracy. So the reason to fix repetitive openers isn’t to game a tool — it’s that the repetition genuinely makes your writing duller to read.

The usual suspects

Three openers do most of the damage.

“The”

There’s nothing wrong with “The.” The trouble starts when it anchors four sentences in a row, because each one then sounds like a label being read off a shelf. “The study found… The results suggest… The implication is… The takeaway here…” Each sentence is fine alone. Stacked, they thud.

“This”

“This” is the sneakiest one, especially when no noun follows it. “This shows the model was unreliable. This means we need new data. This is why the experiment failed.” Two problems at once: the openers all match, and each “This” points back at something fuzzy. What shows it? What means that? Pinning a noun to it (“This inconsistency shows…”) fixes the vagueness and breaks the rhythm in the same stroke.

“It”

“It is important to note.” “It is clear that.” “It should be remembered.” These are filler scaffolds that delay the actual sentence, and they cluster in AI text the way “um” clusters in nervous speech. Most of the time you can delete the “It is” frame entirely and lead with the real subject.

The audit, step by step

Here’s the whole method. It takes one read-through.

1. Open your draft and look only at the first word of each sentence. Ignore everything else. If it helps, highlight just those words.
2. Tally the repeats. You’re looking for runs — three or more sentences in a row with the same opener — and for any single word that starts more than about a third of the sentences in a paragraph.
3. Fix the runs, not every instance. You don’t need fourteen different openers. You need to break the marching rhythm. Changing one sentence in a run of three usually does it.
4. Read the paragraph aloud. Your ear catches what your eye skims. If it still sounds like a metronome, change one more.

That’s it. The discipline is in looking at *only* the first words, because that’s the thing your normal reading brain glides right past.

A before-and-after

Here’s an invented chunk of draft — the numbers are made up to show the pattern — that came out flat and uniform:

> The new onboarding flow reduced signup drops by 12%. This was the biggest gain we saw all quarter. The data also showed faster time-to-activation. This suggests the simpler form was the main driver. It is worth noting that mobile users improved the most.

Look at the openers in order: *The, This, The, This, It.* A perfect alternating drumbeat. Every sentence also runs about the same length, which makes the sameness worse.

Now, audited:

> Our new onboarding flow cut signup drops by 12% — the biggest gain all quarter. Activation got faster, too. Why? The simpler form looks like the main driver, and mobile users improved most of all.

Same (made-up) facts. But the openers now run *Our, Activation, Why, The* — and one of them is a one-word question. A couple of sentences got shorter, one got a dash, and the “It is worth noting” scaffold is gone entirely. It reads like a person reporting a result instead of a system enumerating fields.

A menu of openers to reach for

When you need to break a run, here’s what to swap in. None of these is fancy; that’s the point.

• The actual subject’s name. Instead of “The committee decided,” try “Reviewers split on it.” Naming the real actor is almost always more concrete than a “The.”
• A short time or place phrase. “By Friday, the queue was empty.” “On the second pass, it held.”
• A question. “Why did it break?” A genuine question resets the rhythm and pulls the reader in.
• A dependent clause. “Because the cache never cleared, requests piled up.” This also varies sentence length, which compounds the effect.
• A blunt short sentence. “It failed. Badly.” Two- and three-word sentences are a jolt of human texture that AI drafts rarely produce on their own.

The goal is a mix. If you swap every “The” for an “-ing” phrase, you’ve just traded one tic for another. Range beats any single replacement.

A quick mini-scenario

Say you’re a student polishing a discussion post the night before it’s due. You paste your draft and scan the left edge:

> The article argues that remote work hurts collaboration. The author cites three studies. The evidence, however, is thin. The sample sizes are small.

Four “The”s straight down the margin. You don’t rewrite the ideas — you just touch the openers:

> The article argues that remote work hurts collaboration, citing three studies. That evidence is thinner than it looks. Sample sizes are small, and two of the three studies predate the pandemic.

The argument is identical and arguably sharper, but the opening words now read *The, That, Sample* and the sentences breathe at different lengths. Five minutes, no new research. If you want to go deeper on rhythm and structure, there are more writing-craft guides on the blog that build on this idea, and you can see what PaperBleach costs if a tool fits your workflow.

Frequently asked questions

Is it actually wrong to start a sentence with “The” or “This”? No. Both are perfectly good English, and you’ll start plenty of strong sentences with them. The problem is never one instance — it’s the pattern. When “The” or “This” opens sentence after sentence, the prose develops a metronome quality that reads as generated, because that’s exactly the rhythm many AI drafts fall into. Audit for the repetition, not the individual word.

How many sentences in a row is too many with the same opener? A loose rule: three of the same opener in a row, or the same word starting more than a third of the sentences in a paragraph, is worth fixing. Two in a row can even be a deliberate effect. The point isn’t to hit a quota — it’s to catch the runs where every line marches off with the same first word.

Won’t varying my openers just make the writing feel forced? It can, if you reach for constructions you’d never normally use. The goal is range, not flourish. Most fixes should be small and plain: swap a “The” for the actual subject, drop in a short time phrase, or break a sentence in two. If a reworked opener sounds like a thesaurus wrote it, put it back and find a simpler change.

Why do AI models repeat the same sentence starters so much? Models pick each next word by probability, and at the start of a sentence the safest bets are often the same handful — “The,” “This,” “It,” “In,” plus participle phrases. With nothing pushing them toward variety, they keep returning to those defaults. That low variation in openings is something both readers and detectors can pick up on, though detectors report a likelihood, not a verdict.

Does fixing openers alone make my writing pass AI detection? No single edit does that. Opener variety is one ingredient of natural rhythm, alongside sentence-length variation, concrete detail, and real voice. It helps writing read as human, but detectors weigh many signals and output a probability. Treat the audit as a way to write better, not as a trick to beat a tool.

The takeaway

Repeated sentence openers are a small habit with an outsized effect. They flatten your rhythm, blur your references, and hand a reader the easiest possible clue that something was generated. The cure is just as small: look at your first words, find the runs, and break a few of them with a name, a question, or a short blunt line.

If you’ve got a draft that feels uniform — maybe an AI tool helped you start it — and you want to hear how the rhythm lands before you turn it in, run it through PaperBleach and check it.