I'm a grad student, aiming for a PhD and a career in research in machine learning on encrypted data. That being said, I have experience working in machine learning, but it's my first time working on cryptography and encryption. That being said, whenever I read papers for my project, I can understand the machine learning aspects of it easily (relatively speaking), but I face problems in understanding the encryption parts of it, especially the mathematical concepts and terminology used.

So, to get a 'working' understanding of those topics when I'm reading and interpreting papers, I basically ask Claude and/or ChatGPT to explain me the whys and hows of those. However, I'm a little afraid if I'm relying on these tools too much, and also if the answers that I get from these tools are correct or not, I hope that they are because those are math topics and I'm sure that the bots have been trained on the data.

Additionally, sometimes I also use those to simplify research paper English into simpler language because my vocabulary in general is not the best.

I just wanted some advice on whether this is a right thing to do or will it lead to bad research habits and practices, and how to reduce my reliance on it. To quantify this reliance as a score, I feel that it's like 5-6/10, where a 1 would mean that I do not use these tools at all, and a 10 would mean that I'm basically asking the bot to essentially 'teach' me the whole paper.