While I love it when my coworkers find ways to be more efficient, I’m starting to see a lot of cases where they are using AI to save their own time, but, unfortunately, end up making more work for others. I’ve talked to colleagues at other companies and they are seeing the same thing.
I think what’s happening is that AI often seems like it’s magically doing your work for you, so it’s easy to forget to make sure it’s actually completed the assignment. Regardless of the cause, it’s another form of AI slop, and like online slop, it’s easy to make, but isn’t especially helpful to those around you.
Here are some things to think about to help you make better use of AI without making your coworkers’ day worse:
Always check the work AI has done. For example, if you are using AI to write some Linear tickets, reread them so you know the AI did a good job. Otherwise you’ll end up submitting tickets that lack basic info like context or priorities. It’s helpful to understand, for example, if this ticket came up because a customer requested it.
Make sure your AI understood the assignment. If someone asks you to contribute some information to a shared table, for example, don’t create a new table with a subset of the information they asked for. Sure AI makes that easy, but you haven’t solved the problem, only made more work for your coworker who’s trying to consolidate information.
Don’t just dump AI information on other people. Ask yourself if it solves the problem and reduces the amount of work for everyone. Instead of copy-pasting, edit or refine the answer so it makes sense in the given context. In short: think of the person who is receiving this information and ask if it will help them.
When communicating, remember you are communicating with humans who might have different context from you or the AI. Thinking this way can sometimes be hard even without adding AI slop into the mix, so be sure to give an extra second’s thought to make sure the person you are communicating with can make use of the information in a way that’s helpful for them. AI can’t read minds to understand who your audience is and what they need (yet).
Be cautious sharing AI plans. I’ve seen Linear tickets where people just dump a Cursor or Claude plan in with no further explanation, or with the innocent-sounding “@jane, what do you think of this?” I love the planning features of AI tools, but asking others to review these plans is asking others to do your work. You’ve basically done nothing except prompt an AI, and copied the results.