’Cognitive Surrender’ Isn’t Just Affecting Employees, But Whole Organizations Too

Researchers at the Wharton School have coined the term "cognitive surrender" to describe the tendency to accept artificial intelligence (AI) outputs without scrutiny.

Now, experts are also warning this phenomenon has implications in the workplace and for whole organizations, particularly around decision-making.

A recent Boston Consulting Group (BCG) report shows most companies still struggle to gain value from AI, pointing to people change management as the solution, not solely focusing on the technology.

When organizations tighten deadlines, route more decisions through AI-first workflows, and indicate that fast results outweigh thinking quality, employees feel the pressure to adapt accordingly.

"We're seeing decisions being made unquestioningly because 'AI suggested it,'" Constanza Gómez, CEO and co-founder of Cualy, an AI-powered communication skills-building platform, told International Business Times.

The agreeability of large language models (LLMs) is also an issue. Ask a generative AI to test your strategy, and it will bring up counterarguments—until you indicate which direction you're leaning. At that point, it will mostly find reasons to justify your instinct.

The research on this as a risk is clear. AI's sycophantic tendencies are concerning, according to experts who are also sounding the alarm on the fact that most organizations' architectures aren't yet built with the right guardrails to keep these tools in check.

"AI will reinforce what you want it to," said Kirsten Moorefield, co-founder of team dynamics platform Cloverleaf, an AI coaching platform to improve team communication and performance.

"Even if it starts out introducing different ideas, the moment you indicate your preference, it tells you it's a great idea. That's not a thinking partner but a robotic affirmation."

This is the organizational consequence: companies have inadvertently handed every employee an infinitely patient and relentlessly agreeable collaborator. More and more, these outputs lack the kind of critique and challenge that produce reliable and sound decisions within companies.

Who's Most at Risk (and Why Leaders Usually Get This Wrong)

The fear of judgement is something many have felt at least once in their professional lives — in fact, a recent Workday report found that 37% of people prefer brainstorming with AI for that very reason.

Junior employees can be especially at risk. As Moorefield points out, "Junior staff lack confidence not just in their experience, but relationally. They don't feel safe enough to push back, ask for help, or field critical feedback."

When you combine that relational fragility with a tool that never challenges them, you've created the perfect conditions for cognitive surrender to become the set habit, according to Moorefield.

Senior leaders, however, are not immune either, although their vulnerability shows up differently.

At the executive level, cognitive surrender looks like a strategy built on uninvestigated, AI-generated analysis.

Take an M&A thesis probed by an algorithm rather than a skeptical, seasoned CFO. The stakes are exponentially higher and the oversight worryingly lower still, because senior leaders might assume their experience makes them critical enough. That's where their vulnerability lies.

What Organizations Can Do

Moorefield's recommendation is to mandate human friction and input prior to AI output becoming a decision.

So, before any significant AI-influenced output becomes a plan, route it to someone who will disagree with it. Hand it to a real person, chosen specifically because they see the problem differently.

"We all already know who in our organization will push back on our ideas," said Moorefield. "Choose the ones who are curious, who are different from you, and take your AI-influenced thinking to them before it becomes a plan."

Focus AI governance on middle managers, too, according to Modo, a company that provides expert guidance for organizations to drive AI tool adoption.

Modo's 2026 State of Enterprise AI Adoption research reveals the most common failure point in enterprise AI transformation is not the front line nor the C-suite, but the management layer in between.

However, most enterprise AI programs are designed for frontline employees or C-suite visibility. That leaves the management layer that actually controls whether humans are in the loop (or out of it) to figure it out for itself, and that's where cracks emerge.

Cualy's Gómez echoes this, emphasizing that communication, decision-making, and reasoning throughout an organization are essential to build a collective perspective alongside AI, not as an afterthought.

"Systematic practice is crucial to reinforcing innately human skills like the ability to evaluate and make reasoned decisions," she said.

The discomfort of early-stage, half-formed thinking with another person is not inefficiency to be optimized away, but where, as Moorefield puts it, critical skills actually develop.

"Curiosity, perspective-taking, negotiation, complex problem-solving—without those skills, we lose the ability to combine customer desire with market trajectory and production reality. This isn't a soft-skills problem, but a hazard to our future growth," said Moorefield.

The individual-level coverage of cognitive surrender is important, but it excludes organizations from the responsibility.

The question that hasn't been asked yet, and the one that will define which companies and teams come out of this era with stronger judgement or weaker, is what leaders are going to do about the environment they built.