Archived: how to write less like a bot

This is a simplified archive of the page at https://blog.zgp.org/how-to-write-less-like-a-bot/

Use this page embed on your own site:

01 Jul 2025

ReadArchived

01 Jul 2025

Wikipedia has a page of AI catchphrases, which are phrases and formatting conventions typical of AI chatbots, such as ChatGPT.

The list isn’t just useful for trying to spot if some text was generated by AI. (That’s a hard task that’s probably a bad idea to try, because you end up blaming too many innocent people. As a large language model, I cannot advise you to rely on automated AI detection services.) What’s really valuable about this list is the examples of ways to improve human-written text.

  • is/stands as/serves as a testament, plays a vital/significant role, underscores its importance, continues to captivate, leaves a lasting impact, watershed moment, key turning point ChatGPT puffs up the importance of the subject matter with reminders that it represents or contributes to a broader topic. Might as well replace these assertions of importance with some useful info.

  • rich cultural heritage, rich history, breathtaking, must-visit, must-see, stunning natural beauty, enduring/lasting legacy, rich cultural tapestry These are all opportunities to get specific and show not tell. Use a memorable fact instead of one of these claims?

  • it’s important to note/remember/consider, it is worth, no discussion would be complete without These should go without saying. Otherwise, why put more wear on your one wild and precious set of carpal tunnels even typing it?

  • on the other hand, moreover, in addition, furthermore More extra verbiage. It might make the AI-generated text flow better to a reviewer, but if the sentence is clear without one of these, it can go.

Related: 3 shell scripts: Kill weasel words, avoid the passive, eliminate duplicates by Matt Might.

  • various, a number of, fairly, and quite Sentences that cut these words out become stronger.

  • very, extremely, several, exceedingly, many, most, few, vast. Students insert lazy words in order to avoid making a quantitative characterization. They give the impression that the author has not yet conducted said characterization.

Includes a script to spot these along with example Makefile lines.

#