There is a lot of talk in the writing community about how artificial intelligence chatbots are going to replace authors someday. The fear is that AI will take real writing jobs and we writers will eventually be forced to live in refrigerator cartons beneath an overpass.
This does not scare me because, as an English major, I know from experience that if you choose the right overpass, it’s not so bad.
To further investigate the matter of AI, I signed up for ChatGPT, which is a robot developed by OpenAI. This language-based bot can, without doing any legwork or actual research, piece together fragments of quasi-true information gleaned from the internet and write incredibly crappy essays. Exactly like us writers.
What frightens us writers, however, is that you can ask ChatGPT to write in the style of ANY author, and the computer will mimic an author’s unique prose and voice.
I decided to test this. I logged in and asked the chatbot a few questions. First I asked the chatbot if it
had feelings.
“No, I am not sentient. I am a machine learning model called GPT-3.5, designed to generate human-like text based on the input I receive. I lack consciousness, self-awareness, and true understanding. My responses are generated based on patterns and information from data I was trained on.”
Next, I asked ChatGPT to write 50 words in the style of Sean Dietrich. Which is me. This is where things got scary. Here is what the chatbot wrote:
“Let me tell y’all a little somethin’ ‘bout a Southern breeze on a lazy afternoon. It's like Grandma's sweet tea, cool and a’comfortin’. Makes you forget your worries and sets your soul a’dancin’. So sit for a spell, listen close, and let that breeze a’carry you to simpler times.”
I had to laugh. Because, for starters, we did not call my grandmother “Grandma,” she was always “Granny.” And Granny’s sweet tea…