Google Button Μake us preferred on Google

Headlines asking “Can you tell if the text you’re reading was written by AI?” have become increasingly common in recent years.

Artificial intelligence has been used extensively in text generation, with algorithms like ChatGPT producing promising results.

However, using AI-generated text in scientific research raises ethical concerns such as transparency, bias, informed consent, privacy, and accountability.

AI-generated text, particularly from ChatGPT, can be used for many purposes: generating research ideas, identifying research gaps in specific topics, summarizing countless published articles, and more.

Researchers should be aware that ChatGPT’s outputs are sometimes misleading, or the information is fabricated.

But how much does any of this really matter? Is there something that should concern us more?

Here’s the argument. Instead of focusing on the ethics of using AI to generate text, perhaps we should focus on how to shape the future of writing so that we preserve cognitive thought in the age of AI.

AI and Writing

A recent New York Times essay urged readers to never write using AI, while The Economist and Wired examined how machine-generated prose can be detected.

The questions are becoming more and more familiar: Was this written by AI? Which expressions give the machine away? Can we still tell human writing apart from synthetic text?

These are understandable questions. But they’re misleading for two reasons.

First, AI detectors are inherently unreliable. Humans and machines draw from the same textual heritage. We learn patterns of expression through education and reading. Large language models acquire them through machine learning. Balanced sentences, rhetorical triads, and historical analogies existed long before ChatGPT, yet they’re increasingly treated as fingerprints of artificial prose.

Jovan Kurbalija, founding director of DiploFoundation, described running his own articles from 1996 through an AI detection platform. He concluded, with complete confidence, that more than half had been generated by AI. Among the suspect passages was a rhythmic triad describing how technology affects diplomacy through geopolitics, the issues diplomats negotiate, and the tools they use. The machine implied that a 30-year-old text had been written using technology that didn’t exist yet.

The second problem matters more. By fixating on whether AI produced a sentence, we avoid asking whether the sentence expresses any genuine thought.

AI’s ability to produce polished prose isn’t in itself the main danger. The danger arises when we hand over questioning, interpretation, and judgment to the machine. A grammatically imperfect paragraph that reflects struggle, doubt, and discovery can be more intellectually valuable than a flawless essay produced without real engagement.

The Education Dilemma

This is especially visible in education. When a machine can produce a student essay in seconds, educators understandably turn to detection: Was AI used? Can the student prove authorship?

But these questions address symptoms, not causes.

If the final essay matters more than the intellectual process that produced it, AI becomes an expected shortcut. Students respond rationally to a system that rewards finished products, accumulated credits, and compliance with formal requirements more than curiosity, revision, and judgment.

Here lies a deeper mismatch. AI is fast; cognition is slow. A machine can produce an answer in seconds. Understanding takes time. So does doubt, revision, and disagreement. In an age obsessed with efficiency, certain seemingly inefficient activities, discussion, mental arithmetic, memorization, apprenticeship, remain valuable precisely because they make our brains work.

The answer, then, isn’t banning AI. It’s redesigning writing and learning around the cognitive abilities we want to preserve in the age of AI.

Three approaches could help.

Solutions for AI, Text, and Cognitive Function

  • Use AI as a sparring partner for ideas, not a ghostwriter. Ask it to challenge an argument, identify a missing perspective, formulate a counterexample, or expose a weak assumption. AI should spark more thinking, not less.
  • Value the process, not just the final product. Students, and journalists too, for that matter, should be able to explain how an argument developed, which assumptions changed, why certain evidence was accepted and other suggestions rejected. The better question isn’t “Did you use AI?” but “What did you do with it, and which thinking remained your own?”
  • AI-assisted writing should be grounded in identifiable sources of knowledge. AI prose doesn’t emerge from nowhere or from some miraculous mechanical intelligence. It’s rooted in an existing body of human knowledge and data. Making that grounding visible reveals those intellectual origins. It also offers important side benefits: biases can be traced back to biased sources, claims can be checked against evidence, and AI systems have less room to hallucinate.

The question is no longer whether AI is writing, but whether people are still thinking.