As an NLP researcher who studies AI bias, I have openly talked about the limitations of LLMs. I have met a lot of educators who are very distressed with students who heavily rely on AI. I would agree with them, reprimanding and lamenting the younger generation’s dependance on AI. I disparage that they only look for an easy way to finish tasks and do not take the hard way of acquiring knowledge. But can I criticize them when I also find myself using ChatGPT all the time?

Today, my advisor told me that I used to write my paper so well until I started using ChatGPT. It was a surprise to me. I don’t know why it was a surprise to me. Whenever I read other people’s writing, I could tell whether they used ChatGPT or not. But I naively did not think that others would recognize the use of ChatGPT in my writing.

In my defense, I did my best to avoid the usage of ChatGPT as long as I could. At first, I doubted the hyped enthusiasm over its performance. I did not, and still do not, agree with closed-source models. I did not want to contribute to the detrimental environmental impact of AI data centers by writing a prompt. I had so many reasons to not use ChatGPT. But once I started using it, I could not stop.

My husband first subscribed to a Grammarly plan. Since he paid the money, I had to make the most of it. So I started using it. Then, I discovered that Grammarly provides ChatGPT-powered sentence improvements and generations. Once I tested it out, there was no turning back. I was using AI for all of my writings.

I told myself that I am just using it to check grammar. Since I am bilingual and my English is not as good as other native English users, I need to check my grammar. I am using AI tools so that my English sentences are grammatically correct, formal, and flawless. I am not using an AI to write texts from scratch. I am just asking it to improve my floppy sentences.

Justifications worked. I started using AI more and more for my writing and foolishly thought that others wouldn’t notice. Writing was so easy with the help of AI. Given very rough first drafts, AI will spit out professional and logically sound essays. I could just copy and paste AI-generated writings and save so much time. I no longer had to go through the torment of trying to come up with one sentence for thirty minutes. But my advisor’s comment was a wake-up call to me. Is it really bad to wrestle with your writing for days? Am I not choosing an easy way over the hard way, a very behavior that I characterized as “gen-z” and reprimanded?

It is so tempting to use AI in my everyday work in the name of efficiency. It takes less time to write emails, research papers, and codes. It is like once you start driving, you cannot go back to taking public transportation. Since I am already addicted to its efficiency and proficiency, I cannot say that I will stop using AI. This writing was written without the help of AI, but only because I had an epiphany today: an epiphany that people can tell that I am using ChatGPT.

Another reason that prompted me to write today without the help of AI was my advisor’s comment that my writing used to be really good. One of my justifications for using AI was my insecurity with my writing. I have been asking ChatGPT to “improve my sentence” because I thought my writing is so unkempt. But her compliment encouraged me to write this piece without getting it revised by an AI editor.

Unfortunately, I do not have a conclusion paragraph with a solution. This was more of a confession and a tryout to see if I can still write without using AI. This exercise made me realize that I can still write something. I feel very insecure about this writing, but maybe it is ok. And based on this experience, I could encourage others to also refrain from using AI for their writing. I could tell them it is better to write your own thoughts and words because everyone can tell when you use AI and a lot of them are sick of it. They prefer to see your flaws rather than AI-generated perfection.