I'm unsure whether I should reply because I get the feeling that you think "beneficial to humanity" looks like someone working at OpenAI and not doing research on, say, anthropology.
I chose arbitrary fields that I suspected most readers of HN wouldn't be in, be familiar with, or necessarily esteem just to make the case more obvious. For example, a Ph.D in dance.
There are lots and lots of PhDs in all kinds of things. You can get a Phd in videogames, food hygiene, librarianship, gender studies, ancient greek, theology.
I am sure they are (mostly) worthwhile, and I'm not knocking any of them.
I don't see how they would, by default, be seen to benefit humanity. Some small number might. A very small number. Possibly an extremely small number.
According to one random website there are over 70,000 new PhDs every year.
That's a lot of assumed benefit to humanity. Or.. is it?
I provide this example of a dance PhD thesis topic:
"Deterritorializing Gender in Sydney's Breakdancing Scene: A B-girl's Experience of B-boying" by Rachel Gunn of "kangaroo" breakdance olympic fame.
Anyone can read a book and write an essay. Even AI can write reasonable essays these days. Is AI advancing the pursuit of knowledge?
And you are missing my point about the wide range of fields to which these novice experts are coming from.
Rachael Gunn is an expert in the field of (break)dance, if I extend your position, and her essay is an expert testimony broadening the field of knowledge?
I think it's entirely possible for PhD (or Masters or anyone) to extend debunked theories, to extend or invent nonsensical positions yet to be debunked, to do research of no value, or merely for value to the student and their advisor as a way to increase citations, career advancement, and so on.
I would argue that this is a more reasonable default assumption. In fact, I think the university system favors this likelihood. Students aren't, to my knowledge, asked to examine the worlds problems and take a stab at them.
They are lead into very specific tracks, possibly by their advisor, possibly by the department, possibly by funding or current trends in the field, to do work around a particular area which may go nowhere close to helping all of humanity.
And a "thesis" is just a a really long essay. It has no special quality or instrinsic value, in my humble opinion.
Does fraud happen? Of course. Do people submit more papers on hot topics because they think they are more likely to be approved? Of course. Are some people plagiarizing or misrepresenting the impact of their work? Of course. But the point is to reduce this, and the specific goal organizationally is to reduce these because they harm the reputation and purity of the idea.
I feel like your concern is similar to someone going “firefighting is not actually intrinsically beneficial for humanity”. When someone tells me they’re a firefighter I generally think that they’re doing something valuable. You can argue, well some of them use it as an excuse to get inside a house so they can loot it. Or others are private firefighters for oil companies, and not the “good” community firefighters that save houses. Maybe some of them are doing it for the money and not because they actually particularly care for saving people. But I would counter that the concept is sound and that obvious abuse is, in theory, supposed to be punished.