I recently came across an interesting Nature article titled “AI and the PhD student: Friend or foe?”—and I found myself nodding along more than once. AI is everywhere in research right now. Many PhD students use it daily—to search literature, summarize papers, or generate code. It saves time and can be incredibly helpful.
But at the same time, there is a growing discomfort and unsureness. Many researchers worry that AI might weaken essential skills like critical thinking and writing. And I think this tension captures it quite well: AI can support us—but it can also become a shortcut that prevents us from actually evolving and learning crucial skills.
This gets reflected in conversations with our clients. Now, there is suddenly a growing demand for something that was not needed before: Workshops that cover critical thinking. A skill that used to develop naturally during a PhD—but now seems to need more explicit attention.
In our workshops that include AI, most participants are actually quite cautious. The discussions are less about whether to use AI, and more about how to use it well. A common question in our How to Read a Paper Workshop is: “What is the best AI tool for finding and reading papers?” And the honest answer is: there is no one tool. It depends on your field, your needs—and what the tool might be missing.
An example: A lot of times when using AI to summarize papers, not all information is captured. Nuances can get lost. Which means: we still carry the responsibility. We still need to double-check, question, and interpret.
And this also applies to writing: AI can be very helpful for structuring or polishing text—but relying on it to write entire sections can make it prone to mistakes, and also it is harder to develop your own voice and thinking.
On top of that, there are also ethical considerations: how we use AI, what data we share, and how transparent we are about its use all play an important role in responsible research.
AI is not 100% error-free. (But then again… neither are we.) But it can definitely help in the reading process – see it as the infinitely-patient companion to help you get through your reading.
So… friend or foe?
Probably both.
AI can help us work more efficiently. But if it starts replacing the thinking part, it becomes a problem. Because a PhD is not about speed—it’s about learning how to think, question, be creative and deal with uncertainty.
A small thought to take with you
If you ever feel unsure about your AI use, ask yourself:
Is this helping me think—or is it thinking for me?
And depending on the answer, you might want to adjust how you use it.
We are all still figuring this out.
And maybe that’s okay.