When I am asked to speak about how to use AI, three ideas are always present in my mind.
The first is the principle that one should “Use AI to help you think. Not think for you.”
The second and third ideas build on that principle and spring from these two quotes, which work in tandem.
“AI will magnify the already great difference in knowledge between the people who are eager to learn and those who aren’t.”
Paul Graham“If you want to learn, AI will be great at helping you learn. If you want to avoid learning? AI is happy to help with that too.”
Zvi Mowshowitz
To put this into a visible picture and tie it to AI usage, I developed the simple 2x2 framework below (Business School professors like me think in 2x2’s - it’s a good way to model the world).
One axis measures the above: whether AI is helping you think, or thinking for you. Are you using AI to master new concepts and ideas and grow? Or are you using it to just get cognitive tasks over with and in doing so, avoid learning?
The other axis measures how actively you use AI tools. It’s split between those that use AI reflexively (Power Users) and those for whom AI is an afterthought (Non-Users/Light Users).
Power Users start each project considering “how can I use AI?”, research new prompts, experiment with new ways to use AI, and keep on top of AI developments. In contrast, Non-Users/Light Users will only use AI occasionally for obvious tasks (e.g., replacing the old Google search).
The four quadrants that emerge tell very different stories about where people are headed.
The Quadrant Profiles
Quadrant 1: The Augmented Thinker
These are the power users who take responsibility for their reasoning and learning and remain in control of how and when they use AI. While they use AI reflexively, they use it as a cognitive amplifier rather than a substitute. They brainstorm with it, pressure-test arguments, simulate counterarguments, and prototype ideas faster than they ever could alone. Nevertheless, the resulting judgment, decisions, and learning remains theirs.
The learning and capabilities of Augmented Thinker compounds and accelerates. They learn faster because AI accelerates feedback loops. They become better strategists because they can evaluate far more options in the same amount of time. They develop what is often termed “centaur capability”: human and AI working as an integrated system, each doing what it does best.
The Augmented Thinker must manage the potential risks of this approach. The need for speed can lead to not plumbing the depths of problems. And if their discipline falters, they may fall for the siren song of the ease that AI provides and their independent thinking can erode. But if the Augmented Thinker is cognizant of these risks, they will capture a disproportionate intellectual and economic advantage in the years ahead.
Quadrant 2: The Cognitive Outsourcer
These are also power users, but the dynamic between the person and AI is the opposite of Augmented Thinker. The Outsourcer enters the prompt, but the AI does the analysis. AI writes their paper. AI drafts their strategy. AI forms their opinion. In the process, the Cognitive Outsourcer fails to gain domain knowledge and will lack the insights struggling to gain knowledge brings.
The Cognitive Outsourcer’s performance looks impressive on the outside, but inside it is a hollow shell. When AI fails, hallucinates, or encounters genuine ambiguity, their work product is faulty because they have stopped building the domain expertise needed to catch the errors. When required to talk in public, their comments are shallow and they are unable to answer any question that requires deep reflection.
The long-term risk for the Outsourcer over time is huge: the inability to make decisions on one’s own, the inability to speak off the top of one’s head, and the loss of any form of originality. Worse, the Outsourcer is susceptible to manipulation by the AI systems they rely on, those algorithms, and the people who control them.
Quadrant 3: The Traditional Thinker
These people use little or no AI. They reject it either out of principle (e.g., AI has “stolen” intellectual capital) or because they used it early on and the output they got back was poor. When Traditional Thinkers do engage with AI, it is for minor tasks with low stakes, such as making a recipe or creating the random image. As they did in the pre-AI era, Traditional Thinkers will continue to rely primarily on books, experience, deep conversation, and their own cognition.
While I believe there is great value in using AI, it is much better to be a Traditional Thinker than a Cognitive Outsourcer because they retain the ability for deep and original thought, and the capacity for sustained independent reasoning. In the world of AI, truly independent minds may become rare and valued precisely because of their ability to provide a different point of view.
However, as AI become more powerful, prevalent, and expected in the workplace, there is a big risk to taking this approach. By not using AI, many Traditional Thinkers will be less productive, creative, and knowledgeable than Augmented Thinkers with the same intellectual capability and initiative. In many companies, they will be seen as unable to keep up with technology and be let go.
Some Traditional Thinkers will survive and even thrive if their job requires a human touch. Others will be left behind, not because they lacked intelligence but because they declined to use the tools.
Quadrant 4: The Passive Drifter
Pericles, the great Athenian statesman, said, “You may not be interested in politics, but politics is interested in you.” The same can be said of AI. You might not be interested in AI, but AI is interested in you.
Passive Drifters may not be interested in AI but AI will still heavily affect them. While Passive Drifters may not use AI much themselves, AI is actively shaping them via social media feeds, AI-generated content, recommendation systems and AI summaries of the news.
As a result, the worldview of the Passive Drifter is increasingly curated by forces outside of their control. They are not choosing to resist AI like the Traditional Thinker. Instead, Passive Drifters are simply floating on the currents of AI, consuming rather than creating, becoming spectators in a world whose decisions are being made by those who control AI.
Unfortunately, the number of Passive Drifters is probably the largest of all four quadrants and they will be the most negatively impacted by AI. Unable to offer intelligent answers like the Augmented Thinker, quick responses like the Cognitive Outsourcer, or unique perspectives like the Traditional Thinker, Passive Drifters will fall further and further behind. As such, there is a danger for society as the large Passive Drifter segment falls behind economically. We risk Passive Drifters becoming politically volatile as they lose economic power and become the target of resentment-peddling influencers fed to them by algorithms.
Four Takeaways
First, the real divide is not just between AI users and non-users. It is also between people who preserve human judgment while using AI and those who surrender their reasoning to it. These two factors will divide the winners and losers in the AI era.
Second, which quadrant one falls in is not permanent. You can move between them not only over time but sometimes within a single day. And there may be times when it is fine to outsource some of your cognitive tasks to AI as long as it does not become routine. If you’re an Augmented Thinker, the risk of falling into a habit of overreliance on AI is high and avoiding it takes discipline. The key question to ask yourself for important tasks is, “Am I using AI to think or am I letting AI think for me?”
Third, Traditional Thinkers and Cognitive Outsourcers fall prey to two opposite sides of a coin: not using AI at all or using AI too much. Although I’d judge that being a Traditional Thinker is better than being a Cognitive Outsourcer, both approaches have downsides over the long term.
Lastly, the Passive Drifter problem is an often invisible one that will only grow over time. The fear that people will let AI reason for them is the problem that gets the press. The unseen risk is that many will live a life of passive consumption fed by AI algorithms, leading them to a continued erosion of their agency and making fewer and fewer conscious and informed decisions.
The Brightest Future
Looking across the four quadrants, it appears that the brightest future belongs to the Augmented Thinker. The ones left behind will be those who refuse to use AI, those who outsource their reasoning to it, or those who simply obey it. The winners will be Augmented Thinkers, the people who can use AI reflexively while still thinking for themselves.
Mark McNeilly is a Professor of the Practice at UNC Kenan-Flagler Business School and chairs the UNC Provost’s AI Committee. He writes on AI, leadership, and strategy at Mimir’s Well on Substack.
This article was written by me and improved with additional input from AI.