AT A GLANCE
Artificial intelligence should not be treated as the enemy of education or the workplace. The real challenge is to make sure people really understand the work they are submitting, whether AI helped produce it or not. Instead of relying on unreliable AI detectors or blanket bans, schools and companies should teach responsible AI use, encourage transparency, and evaluate whether people can explain, defend, and apply what they know.
- AT A GLANCE
- We Are Measuring the Wrong Thing
- Should Schools and Companies Accept That AI Is Here to Stay?
- Why Doesn’t AI Detection Measure Human Understanding?
- The Real Skill Is Not Writing. It Is Thinking.
- How Should Companies Evaluate AI-Assisted Work?
- What Is a Better Alternative to AI Detection?
- How Schools and Companies Can Measure Human Understanding
- Preparing People for the Future
- The Future Still Belongs to Humans
We Are Measuring the Wrong Thing
A few months ago, I decided to run a simple experiment. I uploaded an article that I had written entirely by myself to one of the most popular AI detection tools.
Instead, it said there was a 70% chance the essay was written by artificial intelligence. It was entirely wrong.
That event led me to ask a much bigger question. If AI detection software cannot reliably tell whether a human wrote an article, why are schools, universities, and even companies using it to judge students and employees?
The more I thought about it, the clearer the problem became.
We are measuring the wrong thing.
The question should not be whether AI helped write a document. The inquiry should be whether the individual who submitted it knows what is inside it.
Should Schools and Companies Accept That AI Is Here to Stay?
Whether we like it or not, artificial intelligence is becoming another everyday tool.
People use it to write emails, summarize meetings, translate documents, organize ideas, write software, prepare presentations, and answer questions. It is used to study. By students. It’s used by professionals to work faster. Business owners use it to solve problems.
Pretending that AI doesn’t exist is as unrealistic as pretending people don’t use search engines or cellphones.
Students, in particular, have already embraced it. A 2025 survey by the Higher Education Policy Institute found that 92% of university students in the United Kingdom had used generative AI, a significant increase from the previous year. Regardless of institutional regulation, the technology has already arrived in the classroom.
That makes things more complicated.
People are not being stopped from employing AI anymore.
The challenge is teaching them how to use it properly.
Why Doesn’t AI Detection Measure Human Understanding?
Many schools have responded to AI by banning it or by relying on AI detection software. Many companies have done something similar, discouraging employees from using tools like ChatGPT or requiring reports to pass AI detection tests.
The intention is understandable. Organizations want to protect academic integrity, intellectual property, and professional standards.
Unfortunately, AI detectors are far less reliable than many people assume.
Independent research has shown that commercial detection tools produce inconsistent results and can mistakenly identify human writing as AI-generated. In one widely discussed study involving more than 800 writing samples, average detection accuracy was below 40 percent. Small edits to the text reduced accuracy even further.
That means a student who honestly wrote an assignment could be falsely accused, while someone who carefully edited AI-generated content might go unnoticed.
But even if detectors became perfect tomorrow, they would still fail to answer the most important question.
Did the person actually learn anything?
The Real Skill Is Not Writing. It Is Thinking.
Artificial intelligence can write paragraphs, summarize research, and make persuasive arguments.
What it can’t do is replace human judgment.
It can’t judge the real-world plausibility of a conclusion. It cannot take responsibility for its advice. When a decision depends on experience, ethics or practical knowledge, AI cannot determine which solution is best without human judgment.
Those duties remain the responsibility of humans.
And that’s why I think schools need to go beyond just prompt engineering, and start teaching something much more general: AI literacy.
Students need to learn how to ask good questions, verify information, recognize errors, challenge AI-generated answers, protect confidential data, and understand when they should ignore the machine altogether.
Education has always been about learning how to think critically.
AI simply makes that skill more valuable than ever.
How Should Companies Evaluate AI-Assisted Work?
This is not only a classroom problem. Businesses are asking the same questions.
Should employees use ChatGPT? Should AI-generated reports be allowed? Can AI detectors detect work that is machine generated?
These questions sound reasonable, but they miss the point.
Imagine a consultant delivering a strategic report to a client.
Does it really matter whether AI helped organize the document?
Or does it matter that the consultant understands every recommendation, can explain every assumption, answer difficult questions, and accepts responsibility for the final advice?
Most clients care about the second question.
The same goes for lawyers, accountants, engineers, marketers, researchers and executives.
Knowledge work has never been about typing out every word yourself.
It has always been about understanding the problem and making good decisions.
What Is a Better Alternative to AI Detection?
If AI can help people write, perhaps writing should no longer be the only way we measure learning.
Instead of asking students to submit an essay and hoping software can detect AI, imagine ending every major assignment with a short conversation.
Why did you choose this argument?
What evidence made you change your mind?
What sources did you trust the most?
What would you do differently if you had another week?
A student who truly understands the subject will usually answer these questions with confidence.
Someone who simply copied AI-generated text will struggle very quickly.
Businesses can apply exactly the same principle. Instead of asking whether AI wrote the report, managers should ask employees to explain the thinking behind it.
Understanding becomes visible through conversation.
Software cannot measure that.
The purpose of education is not to prove that a machine was absent. It is to prove that a human mind was present, engaged, and capable of explaining what it learned.
I believe the same principle applies to the workplace.
The purpose of knowledge work is not to prove that AI was absent.
It is to show that people know their job, stand behind their decisions and are ready to defend them.
How Schools and Companies Can Measure Human Understanding
Schools and companies do not need complicated detection systems to determine whether someone understands the work they submitted. A short discussion can often reveal more than an AI detector.
A teacher, manager, or client can use the following questions:
- Ask the person to summarize the work without reading from it.
- Ask which evidence shaped the conclusion.
- Challenge one important assumption.
- Ask how the recommendation would change under different circumstances.
- Ask the person to apply the idea to a new problem.
- Require disclosure when AI played a meaningful role in producing the work.
These questions do not punish people for using AI. They test whether the person remained involved in the thinking process.
Someone who understands the subject should be able to explain the reasoning, defend the main choices, recognize limitations, and adapt the idea when the situation changes.
Whether the work was written with a pen, a keyboard, or an AI assistant becomes secondary once genuine understanding is demonstrated.
Preparing People for the Future
Every major technological change forces society to rethink old habits.
Calculators changed mathematics.
The internet changed research.
Smartphones changed communication.
Artificial intelligence is changing how knowledge is created.
History suggests that trying to stop these changes rarely works. Teaching people to use new technologies wisely usually works much better.
Schools should graduate students who know how to question AI instead of blindly trusting it.
Companies should develop professionals who use AI to become more productive without becoming dependent on it.
Both goals require the same thing.
Better thinkers—not better AI detectors.
The Future Still Belongs to Humans
Artificial intelligence will continue improving. Next year’s systems will almost certainly be more capable than today’s.
That should not frighten us. It should remind us what education and professional development have always been about.
Their purpose has never been to produce essays, reports, or presentations.
Those are only evidence.
The real objective is helping people develop judgment, curiosity, critical thinking, and the ability to solve problems they have never seen before.
Institutions that continue measuring whether AI was present may discover they have been evaluating the wrong thing all along.
Those that begin measuring human understanding instead will prepare people not only for the future of AI, but for the future of work itself.
Higher Education Policy Institute (HEPI). (2025). Student Generative AI Survey 2025.
Evidence that AI use among university students has become widespread (92% in 2025), making blanket bans increasingly unrealistic.
Weber-Wulff, D., et al. (2024). Testing of Detection Tools for AI-Generated Text.
https://arxiv.org/abs/2306.15666
Independent evaluation showing that commercial AI detectors are inconsistent and should not be relied upon as definitive evidence of authorship.
Brynjolfsson, E., Li, D., & Raymond, L. (2023). Generative AI at Work. National Bureau of Economic Research (NBER).
https://www.nber.org/papers/w31161
Demonstrates that generative AI can significantly improve productivity, supporting the argument that organizations should teach responsible AI use rather than simply prohibit it.
