Everyone always talks about AI girlfriends. What about AI boyfriends?

No, the tamagotchi creatures hatch from an egg. This is a cat, completely different.

Again, you're assuming that men and women are using the app in the same way, and that a "like" means the same thing. If you assume that women typically only "like" men they would be open to dating, but men will "like" anyone even vaguely plausible with a plan on filtering out poor matches later in the process, you'd see data like this, but it wouldn't support your conclusion. (Is this happening? No idea; again, you'd need more data.)

Further, and much more importantly, you're looking at data showing how often men and women like potential partners, but you're trying to deduce from it how often men and women are liked by potential partners. It's interesting that women apparently only "like" 3.2% of the men they see, but that does not in any way suggest that only 3.2% of men will be "liked" by a woman.

Consider two hypothetical worlds:

In world 1, each woman is very selective (and is only open to dating 1% of men), but every woman has selected a different 1% to be interested in.

In world 2, women are unselective (and are open to dating 30% of men), but all women have chosen the same 30% of men to pursue.

Obviously both worlds seem to differ signiicantly from our own, and each has some challenges! But fairly obviously despite women in world 1 being 30 times pickier, all men could at least in theory find a partner who wanted to date them, whereas in world 2 the majority of men would never do so. And yet if you replicated the graph you linked for world 1, it would have the red bar for the female line take up 99% of the graph. It's really not showing what you think it is.

(Again, I have to stress: I am not trying to claim I know how dating or attraction works, or what the median experience for using a dating app is actually is like; I am instead pointing out that nobody seems to know this, because we lack data.)

That isn't what your linked graph shows. Slightly different hypothetical:

World 1: Men evaluate women on highly idiosyncratic scales. They "like" women in the top 53% of their personal scale, but it's a different scale for each man. All women are "liked" by 53% of men (but it's a different 53% for each woman). Women also evaluate men on highly idiosyncratic scales. They "like" men in the top 5% of their personal scale, but it's a different scale for each woman. All men are "liked" by 5% of women (but it's a different 5% for each woman). Due to their lower standards, men end up "liking" a lot of women who do not "like" them back, but nonetheless, every man will find a mutual match once in every 100 profiles they look at. Women are more selective, and better at focusing their attention on the type of men who might "like" them back, but overall success rates are broadly similar, and every woman will find a mutual match 1.8 times in every 100 profiles they look at.

World 2: Men rank women on a shared objective measure of attractiveness, and "like" any woman in the top 53% of the scale, so 53% of women are "liked" by all men, and 47% are "liked" by no men. Women also rank men on a shared objective measure of attractiveness, and "like" any man in the top 5% of the scale, so 5% of men are "liked" by all women, and 95% of men are "liked" by no women.

Note that in this model on average, 5% of all "likes" from men to women will be mutual, but 95% of men will "like" women and it will never be mutual while 5% of men will "like" women but it will always be mutual. Similarly, on average, 47% of "likes" from women to men will be mutual, but 47% of women will "like" men and it will never be mutual while 53% of women will "like" men and it will always be mutual. (Further note that since most relationships are monogamous, while a bare majority of women can find a mutual match, an overwhelming majority will not have it progress into a relationship since all women are chasing the same 5% of men.)

World 1 matches the graph perfectly, but describes a situation where merely looking at 100 profiles on Tinder guarantees every man a mutual match no matter how ugly or undesirable you may be. World 2 also matches the graph pretty closely, but dooms the vast majority of men and women to be unable to find a relationship at all.

Neither world seems very similar to our own to me (although I do recognise a specific subgroup of men intuit that world 2 is fairly close!), but the fact of the matter is we can't distinguish between world 1, world 2, or reality based on the graph you linked.