There are important studies on the dangers of writing with Gen AI you should be aware of. Most notably:

  • Interacting with LLMs is gradually standardizing human expressions and thought. We are losing the originality in our writing that is the very expression of our character. [1]
  • Interacting with a biased AI writing assistant shifts people’s attitudes on societal issues. This effect persisted even when writers were told about the bias! [2]
  • Students writing with AI consistently underperformed in brain activity, and got lazier with each subsequent essay. [3]
  • With people turning to AI to help advise and guide them, many fail to notice its sycophantic behavior (tailoring responses to what it thinks the user wants to hear). This paper explores that AI might damage real-world relationships, making human ones seem less satisfying. [4]

And the message I want to relay is brief:

Writing is a creative human expression. Your writing doesn’t need to be perfect, the syntactic quirks and word choices in your writing are what make your writing your own. It’s what makes you, you. Your writing is your own.

Present your findings—discuss them with peers and friends. Get into arguments. Allow yourself to be angry, flattered, happy, bitter. Appreciate human feedback.

Highly opinionated friends can be annoying, but all this emotionally-loaded feedback is crucial to your character development. It can be relaxing to talk with a sycophantic machine; but is avoiding blowing your fuses worth the cost of being less human?

With that being said, it’s not wrong to use AI to assist with minor polishes or spellchecking. But be wary it doesn’t come at the cost of being you. Forcibly conforming to a writing style other than your own is secretly a punishment.

[1]

[2]

[3]

[4]