Think about the habits and traditions in your field, think about the nature of your field and do not hesitate to take responsibility for your own (possibly not that great) ideas.
Now, let me elaborate more:
The question encourages personal opinions for a good reason. Various sources on writing research papers differ vastly, though it seems majority does not favor the first person "I" form. For one of the more serious in computer science not in favour of "I", see e.g., Knuth's Mathematical Writing (pg.4) - although later on, the material also discusses the opposite (pg.62 and 113).
Now to a personal position. I do make use of "I" in some contexts. Namely, when I write a paper as a single author and I did so in my PhD thesis. At the same time, you should have clear rules when to mix it with "we" and how. For the dissertation, I explained those rules very early on in the preface: I use "I" whenever the text speaks about my own decisions and choices I made and is the default voice. It means, that it's me who is to blame for whatever incorrect decisions exposed in the thesis. Only if I can show that there is an external force which would push anybody on my place to take the same route, I would use "we" to mean the (research) community, or humankind. I use "we", whenever the discourse is explanatory, such as an exposition of a proof. Therein, "we" stands for "me and the reader". I also strictly use "we", whenever I speak about an insight, or a result which was produced in a collaboration, such as developed in a joint research paper with somebody else. As a side-effect, since this voice is not the default one, occurrence of such "we" always enforces a citation to the joint work, which is a Good Thing.
My personal opinion also is that third person is very bad writing style, since it
offloads responsibility for the presented results to some external entity. As if it wasn't me who made the stupid decision to push that other guy from the cliff, but the guy was (somehow) pushed from the cliff. In my opinion, "we" solves that problem only a little bit, because now the writer admits a bit of responsibility for the act, but still dilutes it by taking into the game somebody else (either the reader, or the abstract research community). Saying "I did this and that and by doing it I personally found this and that" for me is fully taking responsibility for my results. It's not about bragging, or so. Now in some fields, this might be inappropriate, e.g., in pure mathematics, one studies a problem and is not pushed into any arbitrary decisions (e.g., regarding experimental setup), so a style "we" = "the two of us, you, the reader, and me, the writer" is more appropriate.