{"slug": "why-ai-writing-sounds-fake-and-how-to-fix-it-with-the-writ-skill", "title": "Why AI Writing Sounds Fake (And How to Fix It With the writ Skill)", "summary": "A developer built writ, an open-source, prompt-based self-audit skill that scans AI-generated drafts against Wikipedia's cataloged \"Signs of AI Writing\" heuristics and rewrites them to remove synthetic patterns while preserving technical intent. The tool targets six recurring tells, including trailing -ing clauses, clustered stock vocabulary, negative parallelism, em-dash overuse, invented controversy, and metronomic sentence cadence. The author attributes the style to RLHF training that rewards prose sounding complete, polite, and authoritative.", "body_md": "You can spot AI-generated prose within three sentences.\n\nIt rarely stems from grammatical errors. Large language models understand syntax and subject-verb agreement better than most human writers. The tell is subtler: an unearned earnestness, a predictable rhythm, and an insistence on explaining why everything it mentions is historic, essential, or transformative.\n\nWhen readers encounter this tone, they stop reading. Search engines downrank it. Communities reject it.\n\nOver the past two years, Wikipedia editors cataloged thousands of AI-generated article submissions to isolate exactly what gives synthetic prose away. They compiled their findings into a guide called [*Signs of AI Writing*](https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing).\n\nTo turn those observations into an automated, actionable workflow for developers and writers, I built [**`writ`**](https://github.com/Avinashricky211/writ)—an open-source AI self-audit skill that systematically strips synthetic habits out of machine drafts.\n\nHere is an analysis of why language models write this way, the specific patterns that give them away, and how to audit your text before publishing.\n\nThe single biggest tell in AI text is not a specific word. It is the impulse to explain the importance of a statement instead of simply stating it.\n\nHuman writers trust their audience. If you report that an engineering team migrated from PostgreSQL to ClickHouse and cut query latency by 80%, the reader grasps the impact immediately.\n\nA standard language model cannot resist adding a participatory tail:\n\n*\"...slashing query latency by 80%, highlighting the team's commitment to efficiency and underscoring the transformative shift toward real-time analytics.\"*\n\nNotice what happened. The sentence stopped delivering information and started narrating its own importance. If you delete that final clause, the factual value remains unchanged.\n\nLanguage models behave this way because reinforcement learning from human feedback (RLHF) rewards outputs that sound complete, polite, and authoritative. In practice, that training creates prose that sounds like an anxious corporate press release.\n\nThe [`writ`](https://github.com/Avinashricky211/writ) skill codifies Wikipedia's catalog into a targeted checklist. When auditing drafts, these six patterns appear most frequently:\n\nSentences that end with trailing `-ing` clauses designed to manufacture weight:\n\nCertain words are fine in moderation, but language models cluster them with statistical regularity:\n\nLanguage models frequently reach for rhetorical negative parallelism to fake depth:\n\nAI models lean heavily on em dashes (`—`) to splice thoughts together where human writers naturally use commas, colons, parentheses, or separate sentences.\n\nModels often invent controversy around mundane facts to appear objective:\n\nAI prose tends to settle into a metronomic cadence: sentence of twelve words, followed by sentence of fourteen words, followed by sentence of thirteen words.\n\n`writ` Skill?\n[**`writ`**](https://github.com/Avinashricky211/writ) is an open-source, prompt-based self-audit skill designed for AI coding agents, developers, and technical writers.\n\nInstead of generating text from scratch, `writ` acts as an editorial second pass. It ingests an existing draft, scans the text against the codified Wikipedia heuristics, identifies synthetic patterns, and rewrites the passage while preserving the author's technical intent and data points.\n\n```\n                      ┌──────────────────────┐\n                      │ Raw Draft / AI Output│\n                      └──────────┬───────────┘\n                                 │\n                                 ▼\n                      ┌──────────────────────┐\n                      │  writ Audit Engine   │\n                      │  • Kill tacked clauses│\n                      │  • Strip stock terms │\n                      │  • Break cadence lock│\n                      └──────────┬───────────┘\n                                 │\n                                 ▼\n                      ┌──────────────────────┐\n                      │ Natural Human Prose  │\n                      └──────────────────────┘\n```\n\nTo understand how `writ` works, examine this comparison taken from a technical system overview:\n\n*\"In today's fast-paced technological landscape, caching plays a vital role in backend architecture. By seamlessly storing frequently accessed records in memory, Redis acts as a game-changer for distributed systems, underscoring the critical need for low-latency operations and highlighting the shift toward real-time responsiveness.\"*\n\nNotice the symptoms:\n\n*\"Redis stores hot keys in memory to keep read latencies under two milliseconds. For high-throughput services, that eliminates repetitive queries against the primary database and prevents connection exhaustion during traffic spikes.\"*\n\nThe edited version is shorter, contains specific technical metrics, and provides concrete engineering rationale without posturing.\n\nThe repository is hosted on GitHub at [**github.com/Avinashricky211/writ**](https://github.com/Avinashricky211/writ). You can integrate it in three ways:\n\nClone or add the `writ` directory directly into your agent skills path:\n\n```\ngit clone https://github.com/Avinashricky211/writ.git\n```\n\nWhen prompt-instructing your agent, invoke the skill:\n\n```\nReview this article draft using /writ. Strip all synthetic AI tells, remove dangling significance clauses, and ensure the tone reads as a technical peer speaking to another engineer.\n```\n\nYou can copy `SKILL.md` from the repository and paste it into your system prompt when generating or revising blog posts, release notes, documentation, or newsletters.\n\nAdd `writ` as an automated review step in your CI documentation pipelines to flag PR descriptions and markdown docs that lean too heavily on marketing adjectives.\n\nYou can inspect the rules, contribute heuristics, or download the full skill at [**github.com/Avinashricky211/writ**](https://github.com/Avinashricky211/writ).\n\n*Written by **Yadlapalli Avinash Ricky**, AI Engineer, Author of [The Art of AI Prompts](https://avibook.online), and creator of AviGPT-250M.*", "url": "https://wpnews.pro/news/why-ai-writing-sounds-fake-and-how-to-fix-it-with-the-writ-skill", "canonical_source": "https://dev.to/yadlapalli_avinashricky/why-ai-writing-sounds-fake-and-how-to-fix-it-with-the-writ-skill-14mj", "published_at": "2026-10-06 12:41:29+00:00", "updated_at": "2026-10-06 12:48:35.170603+00:00", "lang": "en", "topics": ["ai-tools", "generative-ai", "large-language-models", "ai-agents", "developer-tools"], "entities": ["writ", "Wikipedia", "Avinashricky211", "GitHub", "PostgreSQL", "ClickHouse"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/why-ai-writing-sounds-fake-and-how-to-fix-it-with-the-writ-skill", "markdown": "https://wpnews.pro/news/why-ai-writing-sounds-fake-and-how-to-fix-it-with-the-writ-skill.md", "text": "https://wpnews.pro/news/why-ai-writing-sounds-fake-and-how-to-fix-it-with-the-writ-skill.txt", "jsonld": "https://wpnews.pro/news/why-ai-writing-sounds-fake-and-how-to-fix-it-with-the-writ-skill.jsonld"}}