AI Made It Easier to Produce Work. Did It Make the Work Better?
People are producing more, but more doesn't automatically mean useful. If you made your part faster by making everyone else's harder, where did the productivity gain go?
You can generate a strategy in seconds now. A presentation in minutes. An entire month of content before lunch.
But when you get through all those words, what actually changed?
What problem got solved? What decision became clearer? What became easier for the person doing the work?
That’s where some of this AI-generated productivity starts to fall apart.
People are producing more, but more doesn’t automatically mean useful. Sometimes it just means everyone else has more to read.
I’m all for using AI at work. Use it to get past a blank page, explore an idea, write code, or make sense of information. There’s real value in having a tool that helps you move faster.
My issue is when generating the output becomes the entire contribution.
Someone asks for a recommendation and gets five pages describing every possible option, with no recommendation. A team needs a plan and receives a polished document full of objectives that nobody owns. A customer needs a clear answer and gets four paragraphs that somehow avoid answering the question.
The formatting looks good. The sentences sound professional. There’s just nothing behind them.
You can see it in production work, too. A UX/UI design looks polished, but the user still can’t figure out where to go. A client brief sounds professional, but swap out the company name and it could belong to any business in any industry. A proposed solution repeats the problem in better language without explaining how to fix it.
Sometimes we’re producing the same boilerplate over and over, changing a few words and calling it customized.
Templates have a purpose. Reusing a structure can save time. But a client brief should reflect that client’s goals and constraints. A design should account for what its users need to accomplish. A solution should explain what will change and why that change will help.
If those details are missing, we’ve filled the space without doing the work.
That’s what makes it AI slop to me. It creates the appearance of progress while leaving the actual thinking for someone else.
Someone still has to check whether the claims are true. Figure out which ideas apply. Identify the missing context. Turn vague suggestions into something the team can actually do.
If you generate a document in thirty seconds and send it to five people who each spend twenty minutes untangling it, where did the productivity gain go?
You made your part faster by making theirs harder.
And because AI can make an unfinished thought sound complete, that handoff isn’t always obvious. A rough draft usually looks like it needs work. An AI-generated draft can arrive with an executive summary, confident language, and a conclusion before anyone has tested the idea underneath it.
Polish can hide how little has been decided.
I’m also a little bitter about the em dash. It was one of my favorite punctuation marks before AI made it feel like a suspect. Now I catch myself noticing every one, especially when it shows up alongside the same recycled phrases and perfectly balanced sentences.
An em dash doesn’t prove someone used AI. But when everything has the same rhythm and nothing says anything specific, you start recognizing the pattern. I don’t want people to stop using punctuation they like. I want them to put enough of their own thinking into the work that punctuation isn’t the most memorable thing about it.
The value you bring to your job still includes judgment. You know the constraints, the customer, the history, and the reasons an obvious solution might not work. Your contribution is choosing what matters and taking responsibility for what you put in front of people.
AI can help you develop that thinking. But if you haven’t questioned the output, checked it, or connected it to the actual problem, you haven’t finished the job.
Before sharing something AI helped create, I think a few questions are worth asking:
- Does this answer what was actually asked?
- What did I verify or contribute?
- Is this specific to the client, user, or problem?
- Can someone make a decision or take action from this?
- Would a shorter version be more useful?
Sometimes the best result is a two-sentence answer. Sometimes it’s a working feature. Sometimes it’s realizing the proposed project doesn’t need to exist.
None of those require a ten-page document to count as meaningful work.
I don’t care whether someone used AI to get there. I care whether what they delivered holds up when another person tries to use it.
If AI helps you produce more, great. Make sure “more” isn’t just becoming everybody else’s problem.