I didn't "trust" OpenAI with anything sensitive to begin with. I use ChatGPT quite heavily, including for research. But if things start pushing into the territory of something that I think has actual potential to be genuinely novel/interesting/publishable, I'm drawing the line. And even ChatGPT will tell you to do that. I was semi-joking with it one day and got into the issue of whether or not OpenAI employees read chat transcripts and it said something to the effect of "you can't really be sure, so it's probably best not to put anything truly sensitive into one of these chats" (paraphrased from memory).
I didn't "trust" OpenAI with anything sensitive to begin with. I use ChatGPT quite heavily, including for research. But if things start pushing into the territory of something that I think has actual potential to be genuinely novel/interesting/publishable, I'm drawing the line. And even ChatGPT will tell you to do that. I was semi-joking with it one day and got into the issue of whether or not OpenAI employees read chat transcripts and it said something to the effect of "you can't really be sure, so it's probably best not to put anything truly sensitive into one of these chats" (paraphrased from memory).