Copy & Paste
What is "sloppypasta," you ask?
The Situation
There’s an online writers’ forum I subscribe to, as a way to stay in the loop regarding job opportunities in LA, major news within the screenwriting community, and the latest chatter on production company mandates. Recently, I came across a post where a scribe had raised a concern and was looking for advice: they suspected their rep of using AI tools to provide feedback on their latest story draft without disclosing the fact and without doing much in the way of actually engaging with the story submission on a genuine level. The scribe was wondering how to broach the subject with their rep in a way that wasn’t immediately accusatory, but at the same time would make clear that they were definitely not OK with such practices.
Other screenwriters quickly responded with empathy and support, echoing dissatisfaction with the idea that a paid rep would outsource key work responsibilities and present automated notes as their own thoughts.
Later that same week, I was reading a newsletter from author and publishing expert Jane Friedman, where she shared with readers like me a new word that seemed to summarize the situation perfectly: “sloppypasta.”
What does it mean?
AI slop + copy & paste = sloppypasta.
A website devoted to the subject defines it more concretely this way: “Verbatim LLM output copy-pasted at someone, unread, unrefined, and unrequested. From slop (low-quality AI-generated content) + copypasta (text copied and pasted, often as a meme, without critical thought). It is considered rude because it asks the recipient to do work the sender did not bother to do themselves.”
Not only is the term snappy and memorable, it makes a point that I think deserves attention.
What It Required
Entities like Claude, ChatGPT, Gemini, and a slew of other gizmos advertise with the promise that AI-integrated workflows result in accelerated, amplified, and articulate deliverables. But in a world where AI promises that it can do anything, where do we draw the line between tasks we still do on our own, vs. tasks we hand over for automation?
In the many nuances of the human experience, each individual will draw that line in a slightly different place than their neighbor. The point here is to A) make a conscious choice to draw that line at all, and B) do so with great consideration, discernment, and intention.
As more people bring AI tools into their workplace, the temptation is real. A typical day at the office requires reviewing a draft, answering a tricky question, and/or providing actionable feedback to a colleague. The tool is right at our fingertips, and it’s fast. It produces something that looks thorough. Using an LLM to look and sound intelligent in an email or a Slack message is as simple as copy and paste.
To be clear, AI use in and of itself isn’t inherently bad or something to shame. The issue on the table today is skipping the step that comes after.
Editorial review of submitted materials, creative feedback for a writers’ groups, and other such activities are tasks that still require a human POV. The person on the receiving end is trusting that the other party legitimately showed up for their part of the compact. They’re not looking for something cursory and surface-level; if they were looking for feedback from a machine they could easily do that themselves!
Instead, they’re looking for the other party’s attention in real time: their lived perspective and innate, individual, unique judgment. When AI stands in for that, especially without disclosure, the output replaces deep engagement with superficiality. And the more we talk about it with others, the more we bring the subject to the forefront in workplace contexts, the less likely people will be prone to slip onto that AI slide of downward engagement.
What Changed in the Language
The website devoted to this topic (Stopsloppypasta.ai) offers a short set of guidelines to avoid serving up the messy dish. Here’s a rundown of what they suggest.
Read the LLM’s output before sharing it. If the sender doesn’t read it, they don’t know whether it’s correct, relevant, or current.
Verify the facts. (Remember that little text at the bottom of all AI chat windows?) AI can make mistakes, and can even push back with staunch confidence when the user replies that the output is wrong!
Distill it down to what actually matters for the recipient at this moment. The sender should take the time and effort to actually consider what part (if any) in the generated text is concretely helpful to the specific need, and summarize it in their own words.
When sharing, disclose specifically how AI was involved, so the recipient knows what was checked and what they may still need to confirm. For example: “I worked with this tool, here’s how, and here’s what I’m handing you.”
The Result
Again, I think it’s the act of talking openly about “sloppypasta” with others that can help prevent it from being served up in greater quantities or becoming the new norm.
While LLMs can be useful when you ask them to point out any hidden assumptions or blind spots in your work, they’re no 1-to-1 replacement for an individual’s emotional IQ, out-of-the-box creativity, or unique, lived-in perspective. These human-centric qualities should be highlighted and celebrated more loudly in the office, in the Zoom call, in the LinkedIn job description.
These are the real gems that deliver intelligent results.
On the Page
In my Reedsy profile where I list my editorial services for potential clients, I explicitly state: “Rest assured that I will not use any AI tools to read, analyze, or write a report for your submitted materials. Your intellectual property will remain entirely confidential and secure.”
Moreover, I wrote about my reasoning for drawing the line at this juncture, in the article “Critiquing the Critic: AI vs Human Feedback,” published in Script Magazine. If you want to go deeper, you can find the full piece at this link.
If you like the idea of receiving creative feedback from someone with lived experience, rather than from a machine, drop me a line. I take on a select number of projects each month.
Parting Food for Thought
You’re invited to consider these questions as you go about your week.
In your own professional work, where have you drawn the line between tasks you still do on our own, vs. tasks you hand over for automation?
Have you shared with someone else the place where you’ve specifically drawn this line?
Do you feel able to explain why you made that particular choice?
If you haven’t shared with someone else where you draw the line for AI involvement, why do you think that is?
