Sometimes, my students reply to my emails with messages that appear AI-generated. This has always annoyed me, but only lately I’ve been able to articulate why. Here are the reasons why using AI to do your correspondence with your supervisor is not a good idea.
- No learning takes place. The point of feedback from your supervisor, in learning sense, is that you take the time to reflect it, digest it, and formulate a defensible response to it (in academia, we have to defend our ideas and doings ALL THE TIME). When you cognitively offload this process to someone or something else, then obviously you don’t learn from the feedback. You might think you do, but unless you go through the “hard” part of processing it, you don’t. Learning takes place in a zone of discomfort, and if you avoid the “pain” of thinking through the feedback, then you almost certainly won’t learn anything and the feedback goes to waste.
- Lack of context. The AI doesn’t know the history: it doesn’t know what we have agreed, what you have done or not done in the past, how well, and what your weaknesses as an evolving researcher are. Your supervisor does. So when you let AI write the response, it considers none of this contextual information but just gives a generic, polite, and void response that often appears to blatantly disregard the whole journey thus far (because that’s how the AI works). So, while the response might seem relevant as a reply to the previous message, it typically looks like “out of place” given the bigger picture of getting your dissertation completed on time.
- Non-combative attitude. The AI will not push back. But you, as a PhD student, sometimes should. You might have legitimate reasons for delays. You might have logical reasons for your study choices that just were not understood by your supervisor. But when you ask AI to reply, it will also cave in. It will apologize and take the blame needlessly much, because it has no “skin in the game” and that’s how it was programmed—-to please the user. You, on the other hand, need to push back when your supervisor is wrong (and that DOES happen). When using AI, this pushback doesn’t happen (or even if you asked the AI to “push back”, here again you’re not learning this vital skill, you’re just outsourcing it).
- Sign of disrespect. “Old school” professors (like myself) expect that you do your own thinking and writing. We write to you in our own words, we read your messages with our own eyes. So, when you don’t respond in kind, this can easily come across as disrespectful. Probably you would not want your supervisor to give you generic AI feedback and let AI do the commenting on their behalf—such feedback would have many issues, but that’s a topic for another post. The main point here is that academic exchange between people is based on the people formulating their own thoughts; this does not work if the other party is essentially an AI bot faking a human.
So, these are the reasons why to avoid using AI when emailing your supervisor. It hampers your thinking and writing skills and disrespects the recipient. Instead, the way to go is genuine reflection, acknowledging the big picture and history of the PhD trajectory, pushing back when needed, and learning to communicate like a human, even when it is more “flawed” by appearance than AI-generated text.