You might not have noticed it at first, but you can’t unsee it now: AI is everywhere at work.

It’s not just frequent discussions about AI’s development, or even the growing requirements to use it–now mandated by about 60% of companies . It’s that traces of AI are all over everything your team produces. Emails are crafted with the same syntax. Deliverables all have the same generic tone. Even LinkedIn posts have telltale signs all over them.

It’s not helping. In fact, it’s hurting. Everything your colleagues produce seems to lack their actual brains and skills. And you don’t know what to do about it.

It’s true that AI can make certain things easier at work. It can summarize virtual meeting transcripts. It can take your detailed notes and turn them into a polished first draft of a report. It can even automate repetitive tasks. In fact, it’s linked to at least a 23% average uplift in productivity .

But AI is supposed to give workers more time to think—not replace thinking altogether.

Awareness of AI slop is becoming ubiquitous. We all know the telltale signs: the em dashes (tragically, for those of us who love them), predictable structure, rhetorical questions. It’s gotten to the point that LinkedIn has added a “report AI slop” feature, allowing users to report posts that seem AI-generated. Yet slop is still prevalent.

One in five say low-quality AI work is overlooked if deadlines are met, and 66% spend at least six hours a week correcting workslop errors .

Here’s what to do if you feel like you’re in a room full of robots who used to be people.

1. Make AI use discussable

Decisions about AI use are usually made at the executive level. This means there may not be much talk about how it’s being used on your team.

Spearhead a meeting specifically to discuss AI usage at work. Set expectations going in that this is a learning opportunity for everyone to share how they’re using it, what they’ve learned from it and what they avoid. This will set up an open conversation where no one feels caught.

Here are some questions to get the conversation started:

  • When do you use AI?
  • When is AI appropriate?
  • What work should never be AI-generated?
  • Should AI be disclosed? If so, what about minor use cases?

While opinions on your team might differ, this conversation will allow you to find where you overlap and where you differ, as well as what work is actually being done by AI.

2. Create team guidelines

Now that you’re on the same page, it’s time to set some boundaries.

Set up another meeting with a goal of developing policies and procedures relating to AI that everyone must adhere to.

Most people aren’t intentionally overusing AI. They just don’t realize that rather than streamlining their work, AI is replacing it. They also might not know other people can tell that they’re overusing it.

But by creating team norms, you’ll set the standard without having to address their AI usage directly. It will feel less like a personal attack this way—and more of a team policy.

Examples of AI guidelines include:

  • No AI usage for developing points for performance reviews or sensitive conversations.
  • No AI usage for feedback on deliverables.
  • AI is okay for copyediting writing-heavy deliverables as long as the AI version is not pasted verbatim—each suggestion should be individually reviewed.
  • AI is okay for meeting summarization and sharing next steps after calls.

Setting a policy before AI usage gets even more ingrained in workflows will catch these habits before it’s too late.

3. Write down your guidelines and have everyone sign off on them

Once you’ve brainstormed your team’s guidelines for AI usage, it’s time to cement them.

Print the document and have everyone sign it, signaling their intent to adhere to the norms of AI use. This forms an agreement between your team—not just a suggestion.

Having everyone sign off does more than formalize the policy. It creates real accountability. In fact, the act of complying in writing leads to higher commitment and follow-through.

4. Create a mechanism for reporting AI abuse

Agreement on paper doesn’t automatically mean adherence.

Keep the conversation going by establishing mechanisms to discuss AI use on an ongoing basis.

Start by creating an anonymous form to report AI usage that may be too aggressive.

This isn’t a venue for team members to “tell on” others. It’s a place to bring legitimate concerns that AI slop is repeatedly being produced, so the worker involved can get support. Maybe they have too much on their plate and working with AI is the only way they’re getting through their to-do list. Or maybe they need one-on-one help learning how to work with AI, rather than letting AI work for them. Either way, they need support and the form can help them get it.

You can also create an ongoing segment in team meetings where your colleagues can share use cases where AI helped or hurt more than anticipated as case studies. You can discuss what went well, what didn’t and what to know going forward so your entire team can learn from the experience.

AI, in particular, is increasing its capability and adoption at historic speed, so revisiting once every two years—or even once a year—isn’t going to cut it. Your team will need to adapt with the tech.

Add revisiting the AI policy to your team meeting agenda quarterly. This doesn’t have to be a meeting-dominating topic—just a 15 minute check-in to reevaluate:

  • What AI developments have been released at your company
  • How your team is currently using AI
  • What your team thinks AI should and should not be used for

Update the policy as needed, and ensure everyone re-signs, renewing their commitment.

What if Your Boss is Producing AI Slop?

It’s more common than you’d think. More than half of employees have received workslop from their own manager —and 85% report a loss in faith in leadership as a result.

Managing these situations is delicate—but important. In order to properly support your development, you need to hear from your manager directly, rather than a copy/paste output from their AI model of choice.

Don’t say, “Can you stop using AI?” explicitly. Instead try:

  • "I really value the specific feedback you usually give me. Lately the feedback has felt more general, and I'm finding it harder to know exactly what to improve."
  • "I'd love more examples from your own perspective and personal experience."
  • "Could we spend a few minutes talking through this in person instead?”

The point that you want to hit home is that you want their brain and their experience supporting your work, rather than a generic response output. This respectfully nudges them toward reviewing your work themself without a direct callout.

The workplaces that make it in the AI era won’t just be throwing everything into AI and hoping it works. They’ll be the ones who know the difference between what AI should touch and what it shouldn’t.