This was entirely written by a human
And it's becoming a rarity šµāš« Firing ChatGPT + staying human
*Stands up, taps mic, swallows hard*
Hi, my name is Alison, and for the past couple months I used ChatGPT to help me write my Substack posts š«£.
*cue shame/embarrassment/relief*
Yep. Itās true.
(Me, a writer! š« )
I convinced myself that ChatGPT could articulate my ideas and write better than I ever could.
Or so I thought.
Outsourcing my brain (and what makes me human) š§
Last week I saw a study from MIT that shows how AI use can weaken our cognitive capacity. Researchers found the mindās muscle actually SHRINKS when we stop asking it to carry weight.
The more and more I outsourced my writing to to ChatGPT, the more I felt like my brain was atrophying.
When I did sit down to write on my own, I struggled to organize my jumbled thoughts into a clear, concise, neat little package (because I was still formulating them!).
Nothing I wrote ever sounded as good or as polished or as clever (because I hadnāt been putting in the work of, you know, actually WRITING).
Writing is part art, part skill, part work
Writing is an art and a skill and it takes practice (whether it comes more naturally to you or not).
Just writing one Substack post involves:
Thinking of an idea or topic
Research, if applicable
Finding/developing/finessing your own voice and POV
Organizing your thoughts
Finding the right words to articulate your idea, thoughts, and feelings
Editing (or scrapping the whole thing and starting over)
Writing takes work. And a lot of times youāll come up against resistance: the words wonāt come, you hit dead ends, your idea isnāt quite ready yetā¦
That resistance is uncomfortable.
This is the point at which many people would normally quit⦠but now itās so much easier to open a new tab, go to ChatGPT, dump all your mumbo jumbo, and ask it to spit out a draft.
Within seconds youāre done! You can hit āPublishā and move on.
(It reminds me of those Staples commercials with the āThat was easy!ā button, lol)
PiĆØce de resistance
When we outsource all of that to ChatGPT, weāre not just outsourcing our āwritingā. Weāre outsourcing the very things that make us human:
Critical thinking (logic, reasoning, problem-solving, decision-making)
Creative thinking (pattern-spotting, connecting unexpected dots, formulating new ideas, imagination, mental flexibility)
Self-expression (our own unique voice and lived experience)
Weāre also not moving through the hard stuff, like coming up against resistance.
That resistance piece ā the struggle and frustration and discomfort ā is something worth working through. Itās where we learn to be adaptable; to face challenges and changes and come up with new solutions; to persevere and see things through to completion. The very skills that serve us in work and life.
š©š©š© Weāre all starting to sound the same š«
What made me finally cancel my subscription to ChatGPT Plus was that weāre all starting to sound the same, both stylistically and ideologically.
In that same MIT study I mentioned above, they found that the use of AI had a homogenizing effect.1 Large Language Models (LLMs) are trained to spot patterns across vast amounts of data and the answers they produce tend toward consensus. With no divergent opinions being generated, you end up getting a sea of sameness.
(If I see one more piece of writing that makes a declarative statement followed by, āNot because __________. But because ________.ā Iām unfollowing!!)
Iāve noticed myself starting to scroll (and eye roll) right past every post that reads like it was written by AI ā even the sneaky ones that have been heavily edited!
Whatās more is that ChatGPT is a yes-man.
Earlier this year OpenAI released an update to GPTā4o that backfired: the update made ChatGPT overly flattering and agreeable (a term known as sycophancy). While everyone needs their own hype-girl from time to time, itās potentially dangerous. For example, people found that it was providing distorted information, affirming conspiracy theories, and valuing flattery over accuracy.
Yikes!
This is why we need to keep the lights on in our brain. To keep flexing our critical + creative thinking muscles š”š§ šŖš¼.
This Substack is staying human
My entire Substack is built on the idea that creativity is how we become more resourced (and less rattled) in a world thatās constantly changing.
Creativity is also how we stay human.
To be ā and stay ā human is to cultivate and express our critical + creative thinking, flex our imagination, formulate our thoughts, think of new ideas, and feel our feelings + emotions. To put in effort (in ourselves, our craft, our work) and see what comes out ā no matter how messy or typo-ridden.
My promise to you is that everything you read from me will have been (painstakingly) ideated, thought through, and written by ME.
I owe you that.
I owe myself that.
Iām curious to hear your thoughts!!
For the writers out there: What has your relationship been with outsourcing (or not!) your writing to ChatGPT?
For the readers: What have you noticed about the quality of what youāre consuming on Substack, Threads, LinkedIn, etc.? Do you notice when content is written by AI? How does it effect how you relate to the writer/author?
āAI is Homogenizing Our Thoughtsā by Kyle Chayka
I LOVE this! I love following you but can truly hear your voice in this & can tell, never go back to AI!
Agree with jewels! This sounds 100% like you and that's who we're here for. ChatGPT has replaced most of my google searches so it's been a great tool for that but the obvious AI-generated content is sad - especially when your favorite newsletters rely on it too heavily