Microsoft, MSFT, has published research showing an erosion of critical-thinking skills among workers using generative AI tools such as ChatGPT

Microsoft’s AI Gamble: Innovation vs. Cognitive Decline

Microsoft Corp. plans to spend $80 billion on artificial intelligence this year, solidifying its status as a dominant AI powerhouse. So why did the company just publish a research paper warning about the erosion of critical-thinking skills among workers who use generative AI tools like ChatGPT?

If we’re being charitable, we could say this is just genuine scientific inquiry. But more likely, Microsoft is playing the long game—staying ahead of AI-driven job disruptions while ensuring its tools remain indispensable to businesses. In a tech world fixated on scaling AI models, this kind of self-reflection is a rare and strategic move—one that acknowledges both economic realities and social consequences.

The Study: AI and Cognitive Atrophy

Conducted in collaboration with Carnegie Mellon University, the study surveyed 319 knowledge workers to examine how they integrate AI into their workflow. Respondents included a teacher using DALL-E 2 to generate images for a lesson on handwashing and a commodities trader leveraging ChatGPT to devise market strategies.

A clear trend emerged: The more participants relied on AI, the less they exercised their own critical-thinking skills—including writing, analysis, and evaluation. Many reported a decline in confidence in their abilities, automatically deferring to AI rather than verifying its output.

Time constraints exacerbated the issue. One anonymous participant admitted:

“In sales, I must reach a certain quota daily or risk losing my job. Ergo, I use AI to save time and don’t have much room to ponder over the result.”

A separate study by Anthropic, which analyzed interactions with its AI model Claude, found that its most demonstrated skill was critical thinking—an ironic contrast to the humans using it.

A Future of AI Managers, Not Creators?

The implications are stark: as AI models improve, knowledge workers may shift from content creators to AI supervisors—evaluating AI-generated work rather than producing original ideas themselves. OpenAI’s latest Deep Research model, priced at $200 per month, already scours the internet—analyzing images, PDFs, and text—to generate detailed reports with citations.

According to a Feb. 12 investor note from Deutsche Bank AG, this transformation is happening fast:

“Humans will be rewarded for asking their AI agent the right questions, in the right way, and then using their judgment to assess and iterate on the answers. Much of the rest of the cognitive process will be offloaded.”

Are We Outsourcing Too Much Thinking?

This concern isn’t new. Socrates feared that writing would erode memory. Calculators were once thought to weaken math skills. GPS supposedly destroyed our ability to navigate. And while it’s true that many of us would be lost without our phones, humans have always found ways to redirect mental energy when outsourcing cognitive tasks.

But AI is different. It encroaches on a far broader range of cognition—not just calculations or directions, but decision-making, judgment, and communication. Whether it’s writing a sensitive email or flagging critical issues to a boss, AI now occupies spaces where critical thinking is essential. This raises concerns about professional competency and vulnerability to misinformation.

Why Would Microsoft Publish This?

For a company that profits from OpenAI’s GPT models, this research might seem counterintuitive. But the report itself hints at Microsoft’s motivation:

“If we don’t understand how knowledge workers use AI and how their brains function when doing so, we risk creating products that do not address their real needs.”

Translation? If AI makes employees worse at their jobs, businesses might stop buying it. If a sales manager’s thinking skills decline when using Microsoft’s AI products, the quality of their decisions could suffer—ultimately harming company performance.

Microsoft’s findings don’t just signal concerns about AI’s impact on cognition—they reveal the company’s deeper strategy: adapting AI to keep humans sharp, not just efficient.