Students are delegating the higher-order cognitive skills companies need the most to AI systems and education is making it worse.

The American Federation of Teachers just announced a $23 million partnership with Microsoft, OpenAI, and Anthropic to train educators how to use artificial intelligence "wisely, safely and ethically." While this sounds progressive, research on AI usage in universities suggests we're already creating exactly the wrong kind of future workforce.

When university students interact with AI, they primarily delegate what educators call "higher-order cognitive functions": 39.8% of AI responses involve creating new content, while 30.2% focus on analyzing complex information. Meanwhile, lower-order tasks like remembering facts (1.8%) and basic understanding (10.0%) were rarely delegated to AI.

This represents an "inverted pyramid" of learning that should alarm business leaders. Students are practicing the computational thinking that machines excel at while avoiding the messy, inefficient cognitive work that drives innovation.

Consider what this means for your future hires. Today's college students are learning to approach complex problems by immediately turning to AI for solutions rather than developing their own analytical skills. They're training themselves to be consumers of algorithmic thinking over generators of original insights.

STEM used to be an advantage — Now it might not be

STEM students, particularly in computer science, are early adopters of AI tools. Computer science students account for 36.8% of AI conversations despite representing only 5.4% of U.S. degrees. This might seem positive on the surface because tech-savvy workers are embracing new tools. But dig deeper into how they're using AI, and the picture becomes concerning. Nearly half of all student-AI conversations were "direct" which means that they were seeking answers or content with minimal intellectual engagement. Students are learning to prompt AI systems efficiently rather than developing the problem-solving persistence that breakthrough innovation requires.

When problems don't have predetermined solutions, the kind of challenges that businesses often need to solve to create competitive advantage, these workers will struggle. They've been trained to optimize within existing parameters, not to question whether those parameters need changing.

The business case for human thinking

Companies increasingly recognize that their competitive edge lies not in processing information faster, but in seeing connections that others miss, asking questions that haven't been asked, and adapting when familiar approaches fail.

The Anthropic research shows students are systematically avoiding exactly these capacities. When faced with complex analytical tasks, they turn to AI rather than developing their own pattern recognition abilities. When asked to create original content, they delegate the creative synthesis that is important to industrial innovation.

Across the globe, students are increasingly practicing cognitive dependence on AI during the very years when cultivating intellectual autonomy and resilience is most critical.

Promoters of developing AI in classrooms are fundamentally misunderstanding what students need to succeed in an AI-saturated workplace. Teaching students to use AI tools efficiently is like teaching them to use calculators without understanding mathematics. They become dependent on systems they can't repair, improve, or creatively repurpose when circumstances change.

The most successful companies will not be those with the most AI-efficient employees but those with workers who can combine computational tools with distinctly human insight. This requires cognitive independence, not algorithmic dependence.

What companies actually need

Forward-thinking businesses should demand educational approaches that develop capabilities AI cannot replicate:

• Adaptive problem-solving that transcends algorithmic responses. When systems break down or markets shift unexpectedly, companies need employees who can think beyond established patterns.
• Creative synthesis that connects insights across domains. Innovation happens at the intersection of different fields — exactly the kind of thinking that requires human intuition and cultural understanding.
• Ethical reasoning that guides AI deployment responsibly. As AI systems become more powerful, companies need workers who can recognize their limitations and ensure technology serves human values rather than narrow efficiency metrics.
• Collaborative intelligence that builds knowledge through human relationships. The most valuable insights often emerge from diverse teams thinking together in ways no algorithm can replicate.

If current educational trends persist, there is a growing risk of producing a generation of workers worldwide who can operate efficiently within established systems but struggle to adapt when those systems fail or circumstances shift. These workers may excel at following AI-generated solutions yet be left floundering when those solutions prove inadequate. This creates a vulnerability not just for one nation, but for economies everywhere, particularly as regions such as China, the European Union, and the UAE are actively investing in human capabilities designed to complement AI. Across the globe, students are increasingly practicing cognitive dependence on AI during the very years when cultivating intellectual autonomy and resilience is most critical.

A better investment strategy

Instead of training teachers to use AI more efficiently, that $23 million investment in AI-Teacher efficiency should fund educational approaches that develop irreplaceable human capabilities:

1. Teaching students to ask questions that AI systems haven't been programmed to address. This requires curiosity-driven learning that values wonder over efficiency.
2. Creating collaborative problem-solving experiences that require synthesis across multiple knowledge domains. These challenges should be complex enough to resist algorithmic solutions while building the analytical tenacity that innovation requires.
3. Developing assessment methods that value thinking processes over finished products. As long as education rewards performance over application, students will find ways to outsource the cognitive work that makes learning meaningful.

The future belongs to workers who can combine computational efficiency with human wisdom, moral imagination, and creative synthesis. But this integration requires the ability to guide AI systems rather than becoming dependent on them.

The most successful companies will not be those with the most AI-efficient employees but those with workers who can combine computational tools with distinctly human insight.

Companies that recognize this distinction will actively seek graduates who can think with AI rather than through it. They'll value employees who use artificial intelligence to enhance their capabilities while maintaining the intellectual courage to question its assumptions.

Businesses around the world have a stake in how education evolves. The workforce you will be hiring in the next five years is learning how to think — or outsource their thinking — today. Will they develop the adaptive, creative, and analytical capabilities that fuel competitive advantage across industries and regions, or will organizations find themselves increasingly outsourcing these distinctly human capacities to machines?