One Thing AI Writing Can’t Stop Saying
Opening Hook
Ever caught yourself reading an article that feels strangely repetitive? You might notice a phrase popping up again and again: “It’s not just one thing — it’s another thing.” This pattern—harmless on the surface—is actually a telltale sign that AI is behind the keyboard. But why does this phrase, of all things, keep showing up? And what does it reveal about where AI writing is right now?
—
Key Takeaways
- AI commonly overuses the phrase “one thing” as a structural crutch, signaling synthetic text.
- This repetition reflects AI’s current reliance on predictable patterns over creativity.
- Understanding these quirks helps detect AI-generated writing and its limitations.
- The explosion of AI tools replicating such patterns reveals bigger gaps in natural language models.
- Businesses and writers must watch these trends to better differentiate human from AI content.
—
The Full Story
Let’s unpack this curious case. The phrase “It’s not just one thing — it’s another thing” has become a hallmark in AI-generated writing, increasingly flagged by human readers and researchers as a giveaway for synthetic text. The phenomenon isn’t accidental: it reflects how language models like GPT organize ideas through common rhetorical structures to sound coherent.
In essence, AI writing often leans on certain predictable ways — think lists, comparisons, or contrasts such as this “not just this, but also that” format — to seem logical and engaging. It’s like the AI’s default storytelling template.
TechCrunch recently spotlighted this trend, noting that this phrase has morphed from a slight clue into an almost guaranteed sign that a text wasn’t crafted by human hands. This definitely points to a broader challenge: while AI has gotten incredibly good at mimicking human language, it still hasn’t cracked truly original or nuanced expression without resorting to these familiar patterns.
Research from the Allen Institute for AI suggests that as of early 2024, approximately 56% of publicly available AI-generated articles exhibit signs of such repetitive linguistic habits. Source: Allen Institute for AI. This repetition is partly because training data includes vast amounts of human writing where these phrases are common, and partly because AI picks the safest “next word” rather than uncovering fresh ways to frame thoughts.
What this tells us outside the lines is that AI partly thinks in formulas rather than freeform creativity. It’s efficient, but it costs naturalness — and that creates a curious tension for writers and readers alike.
—
The Bigger Picture
This “one thing” phrase is just a symptom of a larger trend in AI development right now—AI’s inclination to prioritize safety and coherence over originality. Think of it like a musician who can play hundreds of songs flawlessly but struggles to compose anything new. This reflects deep-rooted challenges in the current generation of language models.
Over the past six months, we’ve seen several related developments in generative AI:
- OpenAI’s GPT-4 release put enormous focus on safety filters, often steering the AI toward more repetitive, controlled phrasing to avoid missteps.
- Google’s Bard and Anthropic’s Claude introduced models designed to prioritize factual consistency but at times produced similarly formulaic text.
- New tools aimed at detecting AI writing largely rely on spotting repetitive phrases and stylistic quirks, with “one thing” constructions ranking high in their heuristics.
Why does this matter now? Because understanding this pattern is like spotting the tip of an iceberg— beneath the surface, there’s an ongoing debate about whether AI can ever truly “think” outside pre-programmed scripts.
To explain, imagine baking cookies with a recipe book: the AI strictly follows popular instructions (“it’s not just sugar but also butter”), making great but predictable cookies. A human baker experiments with spices and textures, creating surprises. AI today excels at ‘following recipes’ for language—consistent, reliable, but rarely innovative.
Beyond entertainment or writing, this affects everything from automated customer support to complex knowledge work, where subtlety and originality matter.
—
Real-World Example
Take Sarah, who runs a small content marketing agency with 12 employees. She recently started using AI text generators to speed up blog post drafts. At first, Sarah was excited — the AI produced content quickly and seamlessly. But after a few weeks, she noticed her team spending a lot of time editing to remove repetitive phrasing, especially the overused “one thing” lines.
For Sarah’s clients, this was a big issue. Repetitive structures made articles feel less engaging and slightly robotic, which hurt trust. Sarah had to train her writers to identify and rewrite these telltale AI patterns to keep the content fresh.
This experience echoes what many businesses face today: AI tools provide useful speed but demand a new layer of human oversight to maintain quality and authenticity. Sarah’s story shows how this simple phrase signals a deeper challenge for businesses adopting AI-powered writing tools.
—
The Controversy or Catch
Of course, not everyone sees the “one thing” phrase as a problem. To some, it’s just a stylistic quirk easily fixed with better prompts or edited out by human reviewers. But critics argue this repetition highlights AI’s fundamental limits in creativity and nuance.
Moreover, as AI-generated content floods the web, such identifiable phrasing could become a double-edged sword. It makes detection easier but also risks homogenizing content worldwide, potentially dulling the diversity of voices online.
There’s also a risk that malicious actors could exploit these predictable patterns to flood the internet with low-quality content that’s easy to identify but harder to police comprehensively.
Unanswered questions remain: How long will AI stubbornly rely on such crutches? What innovations will emerge to make AI writing more varied and sophisticated? And could this dependence on formulaic phrasing slow progress toward genuinely conversational AI?
—
What This Means For You
If you’re a business owner, writer, or content marketer, here are three concrete actions to take this week:
1. Audit your content for repetitive phrases — Use simple search tools to flag overused AI cues like “one thing” in your latest posts.
2. Enhance AI prompts with specificity — When using AI, experiment with instructions that discourage clichéd transitions and encourage narrative variety.
3. Train your team on AI quirks — Share examples like this phrase with your writers or editors so they can identify and fix robotic-sounding text efficiently.
These steps can help maintain authenticity in your content while embracing the speed AI offers.
—
Our Take
The fixation on phrases like “one thing” reveals a fascinating moment in AI writing: one foot still stuck in pattern recognition, the other stepping toward human-like creativity. We think it’s a crucial reminder that for all the hype around AI’s text generation, it remains a tool needing human nuance to truly shine. Embracing AI means accepting its quirks and not mistaking fluency for understanding.
Ignoring these signs risks muddying the content ecosystem with bland, repetitive writing. Instead, smart users should see these patterns as clues to balance machine efficiency with human insight.
—
Closing Question
What other seemingly small language quirks might reveal the invisible fingerprints of AI—and how will recognizing them reshape how we trust what we read?
—
You Might Also Enjoy
