My most recent blog assignment had me do something I didn’t expect. We were asked to test how well ChatGPT could write about something I actually know intimately. Just yesterday, I signed my union authorization card with Starbucks Workers United, so I chose to have the AI write a blog post about the process of unionizing a Starbucks from a barista’s perspective. I wanted to see if a machine could capture the reality — the emotions, the risk, the conversations, the strategy — of what it actually feels like to organize in a corporate environment like this one.
The original post wasn’t terrible. Structurally, it made sense. It hit most of the steps: talking to coworkers, reaching out to SBWU, filing with the NLRB, and facing union-busting tactics. But that’s about where it stopped. The tone was a little too polite, too clean, and too detached from what it feels like being overworked, disrespected, and retaliated against just for demanding a voice. It didn’t mention how some of us get pulled into 1-on-1 anti-union meetings or how management bribes and manipulates people. It didn’t get the rage or the hope, the false promises.
So I revised it in Suggesting mode on Google Docs. I added hyperlinks to credible sources, rewrote whole sections in my own words, and even added a photo I took of our union pins made by baristas in Nashville. I fact-checked every major claim. Some parts were right, others needed heavy edits. What I learned is this: ChatGPT is a useful jumping-off point, but not a reliable storyteller (especially when the story is about real people taking real risks). There was no possible way for AI to describe what the process looked like for our store and what we felt like.
The scary thing is, this is exactly how misinformation (such as union-busting) can easily be spread. AI provided a generalized and depersonalized version of something that quite literally changed my life for the better. Without my experience, the whole story felt phony. The future of misinformation is uncertain when we take into account the impact AI will have. Despite this, one thing is clear to me: AI has the ability to exponentially increase the spread of misinformation in American media.
Here’s the final version of my Google Doc: https://docs.google.com/document/d/1fho1zpuRcRf4Q0DO9ySNC4b4Z9h-HSd-IMQBgcvUSf0/edit?usp=sharing
And here’s a link to ChatGPT, the tool I used to generate the original text.
Leave a Reply