AI Recognises Race in Medical Images

Huh?

What do you mean, not working? That the AI was randomly choosing the correct race 82% of the time by luck?

I'm confused by what your implying because it would seem to me that the authors went through many steps to try to pinpoint how the AI was doing this identification and how baffling it was to everyone that even with a lot of x-ray information removed (8x8 pixels compared to say 4k), it somehow was still correctly picking the race.

What would this "something else entirely" that you are implying actually be?

The way we interpret race is largely a social construct due to 'apparent' physical differentiation, less so underlying biological or genetic factors.

We end up segregating ourselves for a variety of reasons, in which case groups that are physically very distinguished end up forming almost an ethnic basis.

For example, two groups with varying genetic makeup and maybe a number of non-obvious biological differences, but who otherwise looked identical - would have the similar life experience in terms of their social treatment by other groups.

But irrespective of how a person is socialized - if you're Black, people are going to treat you one way, and if you're White, people are going to treat you a little different. That 'lived experience' differential is a somewhat unavoidable.

The degree of that variability is obviously debatable, but surely it exists to some degree.

I suppose you could make a parallel in ethnicity: a century ago, the difference between a Scottish-American and an English-American would have been apparent by lineage, accent, Church affiliation, and that might have affected relationships, status etc..

Whereas after a few generations of integration, there is definitely 'no' (or not much) difference between those two groups, and no vector for differentiation/discrimination. The historical ethnic situation was a 'social construct'.

That said, some of the argumentation used to promote the idea that there is no genetic basis for race is a little odd, the 'Africans have more genetic variation than other groups combined' is often used, but frankly I do not understand how that doesn't mean there are material differences between them and other groups.

And of course there is no 'hard line' between groups, but there is also no 'hard line' between the Scottish and English, there are many people who have attributes of both cultures, but that doesn't negate the existence of either group.

I think we're a bit oversensitive these days to these issues. Systematic racism exists and we should think about it, but that doesn't mean there's a boogeyman behind every door.

I think in this case it's also worth examining what exactly the AI is finding out, because it may not be just 'bone marrow'.

I'm very aware that I'm a HN novice, but can I ask why my post title was edited? The new title is much less descriptive, and x-rays are different from medical images, after all.

> That the AI was randomly choosing the correct race 82% of the time by luck?

No; as with the article I linked elsewhere in the thread (https://techcrunch.com/2018/12/31/this-clever-ai-hid-data-fr...), that the AI might have found some other indicator, like filenames in the data set, or metadata in the images that included patient name, or differences in the length of patient name (often redacted by black rectangles in x-rays in training data), or any number of other factors.

This happens all the time in science. As another recent example of "whoops, turned out we were measuring the wrong thing", https://en.wikipedia.org/wiki/Faster-than-light_neutrino_ano...

Another example around AI: https://www.vox.com/recode/2019/12/12/20993665/artificial-in...

> One such résumé-screening tool identified being named Jared and having played lacrosse in high school as the best predictors of job performance, as Quartz reported.

Are lacrosse players naturally better workers? Probably not. Are they probably whiter, wealthier, better networks, etc. than the average population? Probably. These sorts of things - as with the 8x8 pixel example - start to point to confounding variables that need to be worked out and accounted for.

> Human beings have a continuum of features that is heavily interlaced amongst all the 'races'.

That doesn't mean race doesn't exist any more than the fact that height is a continuum doesn't mean that short people and tall people don't exist.