Researchers at Duke and in the Proceedings of the National Academy of Sciences found that people who use AI at work face a social penalty. They are judged more negatively on competence and motivation. Duke researcher Jessica Reif put the paradox plainly: “While AI can significantly enhance work performance, using it may damage your professional reputation.”⁴ ⁵
Think about that for a moment. The tool may make the work better, while making the worker look worse.
That is the trap we are building. We are attaching shame to augmentation.
MarketWatch recently described the double standard now facing new graduates. C.J. Masse, a University of Colorado Denver design graduate, said many of his professors would have flunked him for using AI in designs and images. He avoids AI partly because of ethical concerns, yet he worries employers may expect AI mastery. Aidan Jaramillo, a University of North Texas graduate, had a different experience. Professors encouraged him to experiment with AI in certain settings, and now he uses it in a data-science internship.²
One student was trained to be suspicious. The other was trained to be fluent. Which one did college better prepare for the world that actually exists?
The point is not that students should be allowed to do anything. Jaramillo’s school still barred AI for quizzes and exams where memorized knowledge was being tested. That makes sense.² We still need some closed-book exams, handwritten work, oral defenses, live problem-solving, and assignments where the whole point is to see what a student can do alone.
No one wants a surgeon, engineer, lawyer, accountant, or journalist who cannot think when the machine is off.
But there is a large difference between testing unaided competence and treating every use of AI as moral contamination. A student who uses AI to quiz herself before an exam is not cheating. A dyslexic worker who uses AI to make a rough draft readable is not faking intelligence. A young analyst who uses AI to understand a spreadsheet faster is not lazy if the final analysis is accurate and the person can explain it.
We should care less about whether the tool touched the work and more about whether the human being understands the work, verifies the work, and stands behind the work.
The calculator analogy is not perfect, but it is useful. Calculators did not end mathematics. They changed what we expected students to do by hand and what we expected them to understand. The National Council of Teachers of Mathematics says the body of research generally does not show negative outcomes from calculator use in math teaching and learning.⁸ The successful answer was not to ban calculators forever. It was to decide when students needed unaided practice and when the tool could extend learning.