How to spot AI generated text in articles/posts

With the arrival of AI chatbots, it is common to see many writers copy-paste from ChatGPT, and then sign their name.

I personally don't even give these articles ten seconds of my attention—as they most likely don't represent the author's knowledge or views. I prefer original human writing with all its imperfections.

Here is how I spot AI generated text at sight. I have listed 7 clues.

1. The "It is not XXX, but YYY” template is a dead giveaway.
Eg. "It is not a lack of information, but a lack of judgment."
"It is not technology that is failing healthcare, but leadership."

2. Rule of three's, oft-repeated.
We effectively use it in speeches, orations, but in writing, it is a giveaway — especially when it is repeated throughout.
"Effective healthcare reform requires better funding, stronger governance, and smarter technology. It also requires political will, institutional cooperation, and public trust."

3. "X does not do Y, but Z" along with the em-dash separation.
See this sentence
The above is typical A.I bot language.
A human might write: “Pandemics test our scientific ability and preparedness”
"Pandemics don't test our science—they test our systems, our honesty and our preparedness”

4. Symmetrical sentences that appear overly polished.
E.g. "Good policy requires good evidence; bad policy ignores good evidence"
"It is easy to promise reform; it is harder to deliver it."

5. The Sudden Upgrade sign:
Imagine we already know the person's style of writing/skill level. Suddenly they post something that's WAY ABOVE that level. This is especially common on LinkedIn.
E.g., someone suddenly writes:
"This paradigm necessitates a multidisciplinary epistemological framework to interrogate the evolving healthcare landscape."

6. Template-based wordy introductions that seek to impress the less discerning
"In today's rapidly evolving healthcare landscape, innovation plays a critical role in shaping patient outcomes."

7. Ultra-safe, intellectually hollow conclusions that say nothing useful despite all the vocabulary
Eg. "Ultimately, a balanced approach is required"... blah blah
"collaboration among stakeholders” is the need of the hour...
"thoughtful policy frameworks"

Some caveats:
Human writing will be influenced by bot language, just as we subconsciously copy others' speaking and writing styles.

Humans may genuinely write the above sentences occasionally. But if the whole text "smells" of A.I., it is most likely cut & paste from a bot.

Check the person's prior writings. Look up their works from before 2022, when bots became publicly accessible. Many writers show a gradual improvement or change in style, that's okay, it comes from experience. But if there is an abrupt and dramatic upgrade, they are likely copy-pasting and claiming authorship.

People may use AI to shorten or rephrase their original sentences or check grammar, and that's okay with me.

AI is a tool. I believe it should not be used as a ghost writer or proxy author. We used Microsoft Word, Google docs and thesaurus in the past to check for grammatical errors and to polish our writing. AI can be used in a similar manner— but NOT to CREATE something that we did not intellectually create. If AI was used to create/generate the article, it must be disclosed. The author can't ethically claim full authorship for machine-generated articles.