Don’t Outsource Your Reading to AI

One session I joined focused on the use of AI in research and how supervisors should guide PhD students in this changing situation. Several senior software engineering professors were part of the discussion. The discussion was not about rejecting AI. It was not about fear. It was about something more difficult: how to use AI without weakening the very skills that make someone a researcher.

We discussed many possible uses of AI in research. For example, in a literature review, AI may help you construct search strings, refine keywords, organize papers, support thematic classification, or assist in creating initial codes in grounded theory. It may help you describe your results more clearly. It may also improve the structure, flow, and readability of your writing.

These are useful possibilities. I am not against them. In fact, I believe AI can be a helpful companion in academic work when it is used honestly, transparently, and carefully. But this is also where students and researchers need to be careful. Before using AI in research work, they must check what their university policies say about it.

So before using AI, students should ask simple but important questions:

• Am I allowed to use AI here?
• How should I disclose it?
• Which part of this work must remain fully mine?

Unclear policies should not be treated as permission to use AI without reflection. When in doubt, students should discuss AI use with their supervisors, document how tools were used, and remain honest about what was generated, checked, rewritten, or interpreted by themselves.

For me, one part is very clear.

Do not delegate reading to AI.

AI can help you manage the process around reading, but it should not become the reader on your behalf.

Reading is not only about knowing what a paper, book, or article says. Reading is about the slow process through which your mind enters another person’s thinking. You meet their words. You pause. You disagree. You become confused. You return to the paragraph. You imagine examples. You connect one sentence with another idea you read last month. Slowly, you start building your own understanding.

This is why summaries are not enough.

If AI gives you the summary, who is struggling with the argument? If AI tells you the research gap, have you really seen the gap yourself?

This matters for all students, but it matters even more for PhD students.

A PhD is not a collection of papers. It is not a folder full of summaries. It is not only about producing text. A PhD is the slow transformation of a person into an independent thinker.

I understand this field, and I know where I can contribute.

That ability does not come from shortcuts. It comes from deep engagement with ideas.

Reading builds imagination because you create pictures in your mind. It builds patience because not every idea becomes clear immediately. It builds language because you begin to notice how arguments are formed. It builds judgment because you learn to separate strong reasoning from weak reasoning.

This is why reading opens the door to knowledge. An educated person is not someone who has only received information. An educated person has spent time with ideas, struggled with them, questioned them, and transformed them into understanding.

The same point applies to listening.

Listening is not a shortcut to reading either.

This is worth saying clearly because some people now treat listening as a replacement for reading. They listen to podcasts, lectures, videos, or AI-generated explanations and feel they have “covered” the topic.

Listening matters. Sometimes listening is the first door into an idea. It can create interest. It can help you understand the big picture. It can make complex ideas feel more accessible.

But listening is not the same as reading.

When you listen, the speaker controls the rhythm of the argument. When you read, the rhythm belongs to you.

You can stop at one sentence. You can return to a paragraph. You can underline a phrase. You can sit with one difficult idea until it becomes clear.

Reading gives your mind space to work. And this is the part we should protect.

AI summaries, podcasts, lectures, and videos can all support learning. But they should not become a replacement for the direct encounter between your mind and the text.

“I heard about this” → “I understand this.”

So yes, use AI. Listen to good lectures. Watch useful explanations. Ask for help when a concept is difficult. Use every tool that helps you learn better.

But do not use these tools to avoid the intellectual struggle that creates real learning.

Read the papers yourself.

Read the books yourself.

Read enough that you can see patterns, tensions, contradictions, weak arguments, strong evidence, and open questions with your own eyes.

For students, reading is not an old habit from a pre-AI world. For PhD students, it is not just one academic task among many. It is the training ground where imagination, judgment, language, and independent thinking are built.

Don’t outsource your reading to AI.

Because the moment you outsource reading, you may also begin to outsource the very mind you are trying to develop.