The Greatest Threat in the Age of AI May Be Our Own Cognitive Decline

Consider how much of the work between an idea and a finished project can already be delegated.

An AI system can increasingly determine what needs to be done, create a plan, choose the appropriate tools, execute the work, evaluate the results and make corrections. The person directing it may never need to understand most of the decisions made along the way.

That gives individuals access to capabilities that would previously have required years of experience or an entire team. There is enormous value in that.

But it also means that looking at what someone produces tells you less and less about what they actually know.

A person can publish a convincing analysis without understanding the subject. They can build software without understanding the code, create a business strategy without being able to defend its assumptions or complete an unfamiliar task without learning how it was solved.

The result may still be excellent. The person producing it may simply have developed none of the abilities that the same result once required.

As AI improves, what someone can produce tells us less and less about what they can do independently.

For most people, developing a real understanding of something takes time spent working with it.

You get stuck. You try an approach that fails. You realize that your original assumption was wrong. Eventually, you understand why it was wrong, and that changes how you approach the next problem.

Even when the immediate task is forgettable, some of what you learned remains with you.

An AI agent can now go through much of that process on your behalf. You receive the benefit of the completed work, but the understanding created during the process may remain entirely inside the system.

I think we underestimate how much of our intellectual development happened because there was no convenient way around it.

People learned things because their work required them to. They held complicated problems in their heads because nobody else was going to work them out. Curiosity mattered, but so did necessity.

Remove that necessity and a lot begins to depend on whether people choose to make the effort anyway.

There is an optimistic answer: AI takes care of routine work, leaving people free to focus on more demanding problems.

I expect that will be true for plenty of people. Someone who is already capable and curious can use AI to explore unfamiliar subjects, test ideas and attempt projects that would previously have been out of reach.

But I see no reason to assume everyone will use the time that way.

“More demanding work” is also a moving target. As AI systems improve, the planning, judgment and reasoning we expect people to move on to will become easier to delegate as well.

There is no obvious point at which technology automatically leaves us with exactly the right amount of mental effort.

Physical strength offers a useful comparison. Most people in modern society can get through an ordinary day without performing much physically demanding work. If they want to remain strong, they generally have to create that effort deliberately through exercise.

Parts of cognition may become similar.

People may have to deliberately make room for sustained concentration, difficult reasoning, memory, mental calculation or learning something deeply because their jobs and daily routines no longer require much of it.

The difference is that cognitive decline may be much harder to notice.

There may be very little immediate penalty for allowing those abilities to weaken. You could still be productive, earn well and complete increasingly ambitious projects. Your tools would continue improving. From the outside, you might appear more capable every year.

That is what makes this more complicated than the familiar complaint that technology makes people lazy.

The person could be working hard and achieving a great deal. They might simply be exercising fewer of the cognitive abilities that used to be involved in producing those results.

Your output could improve every year while the abilities beneath it quietly weaken.

This question is one of the reasons I have been building Testuno.app.

Testuno.app is a platform for testing and eventually tracking different aspects of human cognitive performance. That includes reasoning, memory, attention, processing speed and knowledge, along with more specific abilities measured through individual tests.

The idea is not that a single score can define someone’s intelligence or value. Human cognition is far too broad for that.

The purpose is to make cognitive performance visible.

We already measure many things we care about. We track physical activity, sleep, heart rate, productivity and financial performance. But most people have very little objective information about how their cognitive abilities are developing over time.

As AI takes over more intellectual work, that missing information becomes more important.

If someone’s work can be produced with extensive assistance, I want to know what that person can still do independently.

How well can they reason through an unfamiliar problem? How much information can they hold and manipulate in memory? How consistently can they maintain attention? How quickly can they process new information? Are those abilities improving, remaining stable or gradually weakening?

Those are the questions Testuno.app is intended to help examine.

Most cognitive tests provide a snapshot. You complete a test, receive a score and compare the result with other people.

That can be useful, but I think the more interesting question in the age of AI will be what happens over time.

Physical fitness is not established by measuring yourself once. You train, recover, test yourself again and observe whether your performance changes.

Cognitive fitness could be approached in a similar way.

Testuno.app can become a place where people establish a baseline across several abilities and return to see how those abilities develop. Short tests can reveal different parts of the picture: memory, reaction time, attention, reasoning, knowledge and other measurable forms of performance.

No individual result tells the whole story. Together, repeated over time, they can make changes easier to see.

This is especially relevant if daily life no longer provides a reliable indication of cognitive ability. A person’s output may improve because their tools improved, even if their independent performance remained unchanged or declined.

That creates a growing difference between assisted capability and personal capability.

Testuno.app is being built to measure the second one.

AI may also create a different kind of divide than the one people usually discuss.

Even if everyone eventually receives access to similarly powerful systems, they will bring very different minds to those systems.

Some people will use AI while continuing to build their own understanding. They will question its conclusions, notice weak assumptions and recognize when a recommendation does not make sense.

Others may become increasingly dependent on the system not only to complete work, but to interpret information and make judgments for them.

Both groups may produce good results.

But they will have very different relationships with the technology.

Having enough independent understanding to challenge an answer, identify a mistake or pursue an idea the system overlooked gives you some basis for deciding when to trust it. Without that understanding, using AI becomes less like working with a tool and more like accepting the output of an authority you cannot evaluate.

The ability to recognize a wrong answer may become more valuable at exactly the moment when producing a plausible answer becomes almost effortless.

I expect AI to create opportunities for intellectual development that would have been unimaginable a generation ago.

I also expect it to make avoiding that development remarkably easy.

Which outcome becomes more common is something we will have to discover. It may vary between people, professions and generations. It may also depend on whether we begin treating cognitive ability as something worth actively maintaining.

That is the part of this transition I want to understand through Testuno.app.

We will have plenty of benchmarks showing that our models are becoming more capable. We will know how quickly they can reason, code, plan and solve problems.

But we also need ways to observe what is happening on the other side of the screen.

Are people using these systems to extend their abilities, or slowly replacing them? Are they gaining more understanding, or merely gaining access to better answers? Are their minds becoming more capable alongside their tools?

Testuno.app is my attempt to help measure that human side of the equation.

Because in the age of AI, the most important question may not be how intelligent our technology becomes. It may be how much of our own intelligence we continue to use.