# The theory before the court URL: https://deadwritingtheory.com/theory/ > Dead Writing Theory: since ChatGPT, the writing on the web has been collapsing into one machine voice. Same words, same rhythm, same punctuation, everywhere. ## The three counts 1. **Same words.** The vocabulary has converged on one set of marketing words. Prosecution: P-1, P-2, P-3. Defence: D-1, D-4, D-6, D-7, D-8, D-9. 2. **Same rhythm.** The sentences have converged on one set of shapes. Prosecution: no exhibits yet. Defence: D-3, D-5. 3. **Same punctuation.** The punctuation has converged on one mark: the em dash. Prosecution: P-4. Defence: D-2. 8 exhibits bear on the theory as a whole (how much writing is machine-made, how alike it is) rather than on one count: P-5, P-6, P-7, P-8, P-9, P-10, D-10, D-11. ## Where the name comes from Here is the theory, stated plainly. Since ChatGPT arrived, the writing on the web has been collapsing into one machine voice: same words, same rhythm, same punctuation, everywhere. Prose, copy, advertising, marketing. All of it drifting towards a single, slightly shiny, faintly confident hum. You might recognise the shape of it. It's a nod to the Dead Internet Theory, the idea that much of the web is now made by machines, for machines. We liked the name so much we borrowed the grammar. But this site takes only the narrower, measurable corner: the writing itself. Forget the bots and the traffic, and whoever's lurking in the comments. Just the words on the page, and whether they've started to sound alike. We are not affiliated with the Dead Internet Theory or anyone behind it. ## The case for the prosecution The prosecution would like to begin with a list of words you already know in your bones. "Seamless". "Unlock". "Streamline". After ChatGPT, these AI-era marketing words nearly doubled across company homepages, from 1.05 to 1.93 uses per 1,000 words. And it wasn't a few loud outliers: the words rose on 83 sites and fell on only 27. That's a pattern, members of the jury, not a fluke. Exhibit P-2 is one word in particular. "Unlock" turned up 4.4 times as often in 2023 to 2026 as in 2019 to 2022, going from 7.1 to 31.3 uses per 100,000 words. Everybody, apparently, had a door to open. Exhibit P-3 is the one that stings. In 2023, the AI-era words overtook the dotcom words ("solutions", "world-class", "click here") for the first time in 27 years. A whole generation of corporate filler was quietly replaced by its successor, and nobody held a funeral. Finally, the punctuation. In 2023 to 2026, 69% of software homepages use an em dash, against 17% of banking and finance homepages. The industry closest to the machines writes most like them. The prosecution rests, and it would like you to notice the dash count in this very sentence. (There isn't one. We checked.) ## The case for the defence The defence will be brief on adjectives and heavy on dates. Start with the timeline. The rise in those marketing words began before ChatGPT: 0.45 uses per 1,000 words in 2012, 1.29 in 2019. You can't blame the newcomer for a drift that was already well under way. It just turned up late and took the credit. Next, the em dash. It was on 3% of homepages in 2000, 27% in 2022 and 38% in 2026. That's a long, steady climb, and most of it happened before ChatGPT existed. Then the famous sentence shapes: "not X but Y", the reflexive list of three, "furthermore". The things people circle in red. They did not increase. They ran at 0.16 per 1,000 words before and 0.11 after. If anything, they dipped. Now widen the lens. Before ChatGPT existed, press releases used AI-isms 3.7 times as often as news did: 1.61 against 0.44 per 1,000 words. The "AI" sentence shapes were most frequent in pre-1928 speeches (2.07 per 1,000 words) and sermons (1.94), against 0.45 in UK news. And 40 of 43 words now called AI-isms were more common in books in 2019 than in 1980. Last, the "delve" story. Said to have come from Nigerian English, it didn't hold up: 3 of 600 Nigerian news pages used "delve", 1 of 600 Kenyan, 0 of 1,000 UK and US. The defence rests. Quietly. With a small, smug cough. ## Every exhibit ### For the prosecution (10 exhibits) - **Exhibit P-1** (for the prosecution, Case No. DWT-001): AI-era marketing words nearly doubled after ChatGPT. **1.05 → 1.93** uses per 1,000 words. Each of 147 company homepages compared with itself, 2019 to 2022 against 2023 to 2026. Up on 83 sites, down on 27. - **Exhibit P-2** (for the prosecution, Case No. DWT-001): "Unlock" appeared 4.4 times as often. **4.4×** 7.1 → 31.3 uses per 100,000 words. Company homepages, 2023 to 2026 against 2019 to 2022. - **Exhibit P-3** (for the prosecution, Case No. DWT-001): In 2023 the AI-era words overtook the dotcom words for the first time in 27 years. **2023** first year. "Seamless", "unlock" and "streamline" passed "solutions", "world-class" and "click here". - **Exhibit P-4** (for the prosecution, Case No. DWT-001): Software homepages are the most converged. **69%** of software homepages use an em dash (2023 to 2026). Against 17% of banking and finance homepages. Software also carries the most AI-era vocabulary: 3.88 uses per 1,000 words, against 0.39 for insurers. - **Exhibit P-5** (for the prosecution, Case No. DWT-003): About a quarter of corporate press releases involved LLM-assisted writing. **24%** of corporate press releases, late 2024. 537,413 corporate press releases, January 2022 to September 2024. A population-level estimate from word distributions, not a per-document detector. - **Exhibit P-6** (for the prosecution, Case No. DWT-003): Financial consumer complaints took on the machine's words too. **18%** of financial consumer complaints, late 2024. Roughly 18% of 687,241 consumer complaints. Also nearly 10% of job postings in small firms and 14% of UN press releases. - **Exhibit P-7** (for the prosecution, Case No. DWT-005): LLM-dominant websites rose steadily. **2.1% → 29.4%** of sites, second half of 2022 → first half of 2025. Roughly 100k sites archived by Common Crawl, classified site by site. A preprint; on pre-ChatGPT sites the classifier flagged only 0.29%. - **Exhibit P-8** (for the prosecution, Case No. DWT-005): Machine-written sites reach the top of how-to searches. **46.6%** of 10,000 how-to searches had one in the top 10. 65.7% for the top 20 results. 78.8% of LLM-dominant sites have a clear financial incentive, against 55.8% of other sites. - **Exhibit P-9** (for the prosecution, Case No. DWT-006): Machine-written articles drew level with human ones. **50.9%** of new English articles primarily AI-generated, Q4 2025. 49.6% against 50.4% human in Q1 2025. Vendor study: 55.4k Common Crawl articles, three commercial detectors averaged. - **Exhibit P-10** (for the prosecution, Case No. DWT-007): Stories written with AI ideas grew more alike. **10.7%** of the total similarity range, with one AI idea. 8.9% with up to five AI ideas. 293 writers in a preregistered experiment; similarity measured with text embeddings. ### For the defence (11 exhibits) - **Exhibit D-1** (for the defence, Case No. DWT-001): The rise started long before ChatGPT. **0.45 → 1.29** uses per 1,000 words, 2012 to 2019. AI-era vocabulary was already climbing three years before ChatGPT launched. - **Exhibit D-2** (for the defence, Case No. DWT-001): The em dash was spreading for two decades. **3% → 27%** of homepages, 2000 to 2022. 38% by 2026. The rise after ChatGPT continues a line that started in 2000. - **Exhibit D-3** (for the defence, Case No. DWT-001): The sentences did not change. **0.16 → 0.11** AI-style sentence shapes per 1,000 words. "Not X but Y", reflexive lists of three, "furthermore": no increase on company homepages after ChatGPT. - **Exhibit D-4** (for the defence, Case No. DWT-002): Press releases wrote like this before any model did. **3.7×** as many AI-isms as news, pre-ChatGPT. 1.61 against 0.44 uses per 1,000 words in pages from 2021 and earlier. - **Exhibit D-5** (for the defence, Case No. DWT-002): The "AI sentence shapes" were most at home in old speeches and sermons. **2.07** per 1,000 words in pre-1928 speeches. Sermons and religion 1.94; UK news 0.45. - **Exhibit D-6** (for the defence, Case No. DWT-002): 40 of 43 "AI-isms" were already rising in books. **40 / 43** more common in 2019 than 1980. Google Books. "Transformative" rose 31 times between 1980 and 2019. - **Exhibit D-7** (for the defence, Case No. DWT-002): The "delve came from Nigerian English" story did not hold up. **3 / 600** Nigerian news pages used "delve" (2021). 1 of 600 Kenyan, 0 of 1,000 UK and US. Rare everywhere; the difference was within chance. - **Exhibit D-8** (for the defence, Case No. DWT-004): Lexical diversity in news did not fall. **214.45 → 254.65** MTLD lexical diversity, 2018 → 2024. Two samples of roughly 30,000 news articles each. Machine-style words did rise, from 0.230% to 0.347% of words, but the homogenisation did not show. - **Exhibit D-9** (for the defence, Case No. DWT-004): Two other diversity measures barely moved. **0.00110** change in MATTR lexical diversity, 2018 → 2024. Against within-year variation of 0.00109 (MATTR 0.88011 → 0.88121). Maas moved from 0.01469 to 0.01482, a difference of 0.00013 against 0.00016. - **Exhibit D-10** (for the defence, Case No. DWT-006): The machine share stopped rising. **49.9%** of new articles primarily AI-generated, Q1 2026. "Since Q1 2025 the percentage of primarily AI-generated articles has plateaued at roughly 50%." Vendor study, detector-based. - **Exhibit D-11** (for the defence, Case No. DWT-007): Each story got better with AI ideas. **8.1%** rise in novelty with up to five AI ideas. Usefulness rose 9.0%. Less creative writers gained most: 10 to 11% on creativity, 22 to 26% on how enjoyable and well written the story is. ## How evidence is admitted Every court needs rules, or it's just a pub argument with better chairs. Ours are short. In our own studies, the questions and the word lists are fixed before anything is measured, so nobody gets to go fishing for a result they like. Everything we measure is measured with deterministic code, never a language model. Same input, same answer, every time, with no vibes involved. Other people's studies are entered too, credited by name, with their methods and caveats read into the record. Every case links to its full method, and to the original study. If you think we got it wrong, you can check, and we'd honestly rather you did. And both sides are always shown. The prosecution and the defence get their full say, even when one of them is clearly having a better afternoon. 1. **The question is fixed first.** The question is written down before anything is measured. Each case states it at the top, like a charge. 2. **Counted by code, never by a model.** Our own studies are counted with deterministic code that gives the same answer every run. No language model is asked for an opinion. A third-party study is entered with its authors' method stated and its caveats beside it. 3. **Every exhibit goes in the record.** Every finding is entered, for the prosecution or for the defence, and the full method and data are published on ScriptGrain, or, for a third-party study, linked at the original. ## The docket so far - [Case No. DWT-001: 27 years of company homepages. The words changed. The sentences did not.](https://deadwritingtheory.com/cases/did-chatgpt-change-how-companies-write/) (split verdict; ScriptGrain SGR-004; updated 2026-10-04): On company homepages the "AI voice" is mostly a vocabulary, and an older one than its name suggests. ChatGPT appears to have sped the drift up rather than started it. Markdown: https://deadwritingtheory.com/cases/did-chatgpt-change-how-companies-write.md - [Case No. DWT-002: Where the AI-isms came from. Not an invention. An inheritance.](https://deadwritingtheory.com/cases/where-did-ai-isms-come-from/) (verdict for the defence; ScriptGrain SGR-007; updated 2026-10-04): The "AI voice" looks less like an invention than an inheritance: the vocabulary of corporate announcements laid over the rhetorical shapes of older speeches and sermons. Markdown: https://deadwritingtheory.com/cases/where-did-ai-isms-come-from.md - [Case No. DWT-003: A quarter of the press release. The machine is in the room. In a lot of rooms.](https://deadwritingtheory.com/cases/a-quarter-of-the-press-release/) (verdict for the prosecution; third-party study: Liang et al., Patterns 6(12), 2025; updated 2026-10-04): By late 2024, roughly 24% of corporate press releases involved LLM-assisted writing. The study measures adoption, not sameness, but the machine is in a lot of rooms. Markdown: https://deadwritingtheory.com/cases/a-quarter-of-the-press-release.md - [Case No. DWT-004: The news that didn't flatten. The fingerprints are there. The flattening is not.](https://deadwritingtheory.com/cases/the-news-that-didnt-flatten/) (verdict for the defence; third-party study: Fitterer et al., Proceedings of ACL, Student Research Workshop, 2025; updated 2026-10-04): Machine-style words rose in English news between 2018 and 2024, but lexical diversity did not fall. A real point for the defence on vocabulary, not a knockout. Markdown: https://deadwritingtheory.com/cases/the-news-that-didnt-flatten.md - [Case No. DWT-005: The websites nobody wrote. More sites with little human input. More of them selling.](https://deadwritingtheory.com/cases/the-websites-nobody-wrote/) (verdict for the prosecution; third-party study: He et al., arXiv preprint 2605.00087, 2026; updated 2026-10-04): LLM-dominant websites rose from 2.1% to 29.4% of sites sampled between late 2022 and early 2025, and they turn up in how-to search results. A preprint, detector-based. Markdown: https://deadwritingtheory.com/cases/the-websites-nobody-wrote.md - [Case No. DWT-006: Half the articles, and a plateau. Half the articles. Then the rise stopped.](https://deadwritingtheory.com/cases/half-the-articles-and-a-plateau/) (split verdict; third-party study, vendor: Paredes et al., Graphite research (Five Percent), 2026; updated 2026-10-04): A vendor study: primarily AI-generated articles reached about half of new English articles by Q1 2025, then the share stopped rising. Detector-based, not peer reviewed. Markdown: https://deadwritingtheory.com/cases/half-the-articles-and-a-plateau.md - [Case No. DWT-007: Better stories, more alike. The stories grew alike. Each one got better.](https://deadwritingtheory.com/cases/better-stories-more-alike/) (split verdict; third-party study: Doshi and Hauser, Science Advances, 2024; updated 2026-10-04): In a preregistered experiment, AI story ideas made each story better and the stories as a group more alike: individually better off, collectively narrower. Markdown: https://deadwritingtheory.com/cases/better-stories-more-alike.md - Case No. DWT-008: pending, under investigation. The next case is being measured. ## Instructions to the jury ### Is this the Dead Internet Theory? Dead Writing Theory is a nod to the Dead Internet Theory: the idea that much of the web is now made by machines, for machines. We take the narrower, measurable corner of it, the writing: prose, copy, advertising and marketing. Not affiliated with the Dead Internet Theory or any site, forum or person associated with it. ### Is this a real court? No. The Court of Public Prose is a figure of speech for how we weigh the evidence: two sides, every exhibit in the record, a verdict per case. No court, agency or government is involved. ### Do you use AI to measure the writing? Not in our own studies: they are measured with deterministic code, never a language model, and the questions are fixed before anything is measured. Third-party studies are entered with their authors' own methods, some of which use AI-text detectors; each case says so in its caveats. ### So is the web’s writing dying? We don’t argue for the theory or against it. So far: 7 cases, three split verdicts, two for the prosecution and two for the defence. The jury is still out. ### Who runs the studies? ScriptGrain runs our own studies, and each of those cases links to its full method and data on scriptgrain.com. Third-party studies are credited to their authors and linked to the original. Ours: made by the same studio that runs this site. ### Can I submit evidence? Yes: hello@deadwritingtheory.com. The next case is being measured.