Is AI flattening your team’s creativity? Here’s how to tell. - Inside Atlassian

Skip to content

Subscribe

Dismiss

Subscribe to our newsletter

Research and insights on how teams can deliver results with AI.

Email Address

Sign up

Thank you for subscribing.

Your time is valuable. We promise to send only what’s actually worth reading.

Search

Is AI flattening your team’s creativity? Here’s how to tell.

May 20, 2026

Published In AI at Work

Think about your to-do list for a given work week. From responding to a colleague’s quick message to building a team strategy, how do you prioritize how much time and energy to allocate to each task?

Behavioral science tells us most of us don’t optimize; we satisfice — we find the first “good enough” option, and we move on. The term, a mash‑up of “satisfy” and “suffice,” was coined by Herbert A. Simon, a mid‑century Nobel Prize–winning economist and sociologist who studied workplace decision making. Satisficing, Simon argued, is a pragmatic way to work, especially for quick-turn, low-stakes tasks or for complex decisions where the complete picture is ambiguous or unknowable.

But while satisficing may have made sense for generations of workers, AI has rewritten the rules. Now, “good enough” appears instantly, often before anyone thinks deeply about the question or challenge at hand. That speed might be great for the small stuff, but when it becomes the norm for everything, teams risk sliding into sameness, shallow reasoning, and — at worst — making the humans in the equation entirely irrelevant.

Atlassian’s view is simple: AI should amplify human judgment, not replace it. But given our natural inclination to satisfice, that won’t happen by default. Teams need to intentionally decide how and when to invite AI into the room, choosing the mode of collaboration that best fits the project at hand.

This post introduces the AI Collaboration Framework : a tool designed by Atlassian’s Teamwork Lab that teams can use to turn the dial from AI‑first to human‑first thinking, depending on the needs of the project. Use it to safeguard the capabilities that make your people valuable: human judgment, taste, and the ability to think deeply and originally.

Why over‑reliance on AI is risky: three evidence‑backed pitfalls

Pitfall 1: AI‑induced groupthink

Good leaders don’t let the loudest voice go first. They use techniques like brainwriting to surface diverse ideas before discussion. But with AI, many teams are doing the opposite: asking the model to go first, then reacting. Research is beginning to quantify the cost of this approach. Studies of human-AI creative writing show individuals may look more creative on average when they use AI, but the set of outputs as a group narrows and clusters around model‑suggested patterns.

When every whiteboard starts with an AI‑generated draft, you get speed at the expense of variety. That’s dangerous for strategy and innovation, where outliers and contrarian angles often drive breakthroughs.

Pitfall 2: “Cognitive surrender” and loss of deep work muscle

Early neuroscience and human-computer interaction research indicates heavy reliance on AI corresponds with lower neural engagement and weaker recall of one’s own reasoning. In one MIT study, LLM users showed the weakest brain response during a writing task and struggled most to accurately quote their own work. Even more concerning, when heavy AI users switched to writing without AI, their neural response remained lower, indicating a lasting decline.

Researchers at Wharton coined the term “cognitive surrender” to describe this effect, finding that when study participants had access to AI they often deferred their own reasoning to it, getting more accurate when the AI was right but missing mistakes when it was wrong. The researchers also found that simply having access to AI inflated participants’ confidence in their accuracy, even on incorrect answers. And this isn’t just a lab effect: in a recent Anthropic survey of 81,000 global AI users, cognitive atrophy was the fourth most cited concern.

Teamwork Lab observed a similar pattern in our qualitative research studies. While AI may make some work feel easier, that convenience can dull your thinking and erode your appetite for tackling the hard problems. “It’s easy to just plug every question you have into AI,” said one software engineer. “When you need to think through something that’s too complex for AI, or you can’t trust AI to do it, then you’re faced with, ‘Oooh, I haven’t done this in a while.’ It feels harder than would have been maybe a year or two ago.”

Pitfall 3: Hollowing out your own value

In a world where anyone can instantly generate a decent draft, the most precious resources are your judgment and taste: your innately human ability to decide what matters, spot what’s missing, and shape a point of view that fits this customer, this market, this moment.

In cognitive science, the “illusion of explanatory depth” describes...