If you’ve spent any time online recently, you’ve probably noticed a trend: more and more articles, blog posts, and social media updates are being drafted by ChatGPT or similar AI writing tools. And while these tools are amazing for productivity, there are some quirks that consistently make readers pause and raise an eyebrow. As someone who reads a lot of AI-assisted content, we’ve compiled the top five ChatGPT pet peeves you’re likely to see in 2026—and why they can be so irritating.
1. Em Dashes That Feel Out of Place
Yes, em dashes can be useful. Yes, they can inject drama or clarify meaning. But when ChatGPT is used to generate text, em dashes suddenly appear everywhere. Mid-sentence, in bullet points, even in headlines. Even when perfectly correct, they can make readers feel like the writing is trying too hard. In 2026, after years of AI content everywhere, we’ve collectively developed a subtle aversion.
2. The “Conclusion” Paragraph
We all know the feeling. You read an article, it is interesting and engaging, and then it hits you: the dreaded conclusion paragraph. ChatGPT loves these. It loves to summarise, wrap up, and restate points neatly. But often, it feels forced, repetitive, or bland. Instead of leaving the reader inspired, it feels like the writer was following instructions from a template. Readers in 2026 want insights that linger, not a neat bow at the end.
3. UK Companies Using American Spelling
This one drives native readers up the wall. A UK-based company publishes a blog post, and suddenly we’re seeing “color” instead of “colour,” “organize” instead of “organise,” and “realize” instead of “realise.” It may seem minor, but it immediately signals that either the writer or the AI hasn’t considered its audience. UK readers notice instantly. Small details like this can undermine credibility.
4. Overuse of Generic Phrases and Clichés
ChatGPT has a habit of generating content that is technically correct but overly generic. Words and phrases like “cutting-edge,” “game-changer,” “in today’s fast-paced world,” and “at the end of the day” show up repeatedly. While not wrong, they make content feel bland or formulaic. Readers are looking for personality, voice, and specificity. AI-generated clichés rarely deliver that.
5. Lists That Go On Forever
AI loves lists. Ten tips, seven reasons, five steps. Lists are great for scannability, but ChatGPT sometimes goes overboard, padding content just to reach a word count. You end up with 15 points where 5 would have sufficed, or steps that are vague and repetitive. Overlong lists frustrate readers looking for concise, actionable advice.
Honourable Mentions
Other quirks: ChatGPT favours neutral or overly polite language, which can make writing feel bland. It sometimes repeats phrases or sentences unnecessarily, especially when asked to “expand” content. It occasionally includes filler like “it is important to note” or “as previously mentioned,” which can feel redundant.
The Takeaway – Yes this last paragraph or we could title it ‘Conclusion’
These pet peeves aren’t the fault of ChatGPT. It’s an incredible tool. The problem arises when AI is used without a critical eye. Em dashes that jar, forced conclusions, incorrect spelling, clichéd phrases, and overlong lists are all signs a human edit is needed.
Readers notice these patterns, even if they cannot always articulate why something feels slightly off. As AI becomes more common in writing in 2026, recognising and correcting these quirks separates good content from mediocre content.
At the end of the day, ChatGPT can save time, but editing for style, voice, and audience is more important than ever. Pay attention to em dashes, skip the obvious conclusion, use local spelling, avoid clichés, and keep lists concise. Do that, and you’ll create articles that feel both human and professional, even in 2026.