Hacker News Reader

Comment by mustafa_pasi | Hacker News Reader

That is not my understanding of the ethical dilemma. People are different, and belonging to groups is highly predictive, and that is what ML algorithms depend upon to classify you (every feature is a group). The ethical dilemma is, should you be doomed to be assessed based on your belonging to an arbitrary set of groups? It is highly predictive, but also by definition, discriminatory.

Business entities benefit, because on average the predictions are right, but individuals are being discriminated against, and btw, besides the protected categories, other types of discrimination are also unfair.

View on HN

You should be allowed to think and perceive and notice things, and use tools to help you do so. It is ridiculous to mandate algorithms should be modified to hide statistical facts about the world.

by all means, educate people that just because some population is perhaps on average more criminal, it doesn't imply any individual belonging to that population is also more criminal.

To hide the statistics would also hurt efforts to find solutions. That can not be a good thing.

Edit: if you don't like the criminal example, what about noticing some populations are on average poorer than others? Would that be a good thing, because government money could be diverted their way, or a bad thing, because banks would be less inclined to lend them money, and landlords would be less likely to rent them a home?

The point of contention is not whether ugly statistical data should be hidden away. That is a stupid conspiracy theory. Sociologists have no problem publishing all kinds of ugly statistical data. The problems start when for profit entities or governments, start deciding to discriminate against individuals because they belong to certain groups. Statistical data applied to individuals is always going to result in injustice. The thought of the economy at large, or the government reducing an individual to a few datapoints and deciding their fate based on group statistics is just about the scariest communist dystopian nightmare, as far as I am concerned. It robs people of their own humanity.

You can't just call everything you don't like a conspiracy theory. I have seen plenty of people on HN arguing for censorship of unflattering statistics to achieve social justice goals. Why not just work on fixing the causes of the unflattering statistics in real life, rather than inventing a new framework to make ML models provide less accurate predictions? How do you know these manual override techniques aren't introducing inequity somewhere else in the model? Pareto optimality seems like it should be a very real concern here.

> I have seen plenty of people on HN arguing for censorship of unflattering statistics to achieve social justice goals.

I've never seen this. Can you link some examples?

I invite you to spend lots of time digging through search results, which I myself decline to do since these conversations are rarely productive, however this frequently comes up whenever there's a thread discussing declining white demographics in the US. There is usually someone who says we should stop collecting these statistics altogether because it gives fuel to white nationalists. There are more examples beyond that, but those stick out in my mind because they were so blatant.

And while people will surely respond to this post with "Ha, got him! He can't link to a specific example!", I don't believe it's a reasonable expectation of me to catalogue the thousands of comments I read on this website.

I'm interested in those threads (I do a lot of work with demographics), but I've never noticed that.

Searching for "US demographics collection nationalists" finds one result (your comment above - and presumably soon this one): https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

I tried "US data collection nationalists" but that only found one page of results none of which supported your assertion: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...

Do you have a better set of search terms?

HN's search feature leaves a lot to be desired when it comes to comments. Perhaps some variation of "replacement conspiracy theory".

Fwiw, I couldn't find anything resembling what you're saying in the last 2 years under "great replacement" and similar terms.

I've tried "white replacement data" and "white replacement stats" and "white replacement statistics" (without quote marks) with no luck in the last 2 years.

The closest I've found is [1] which is a downvoted comment saying:

> Census data indicate White population has shrunk since 2010

Why is this the first bullet point? Why a focus on White people? Why do you care HN?

The only other place I see stuff like this is when it’s used as an alt-right talking point, or by white-supremacist mass-shooters who write about ‘The Great Replacement’ [1]

So that isn't even arguing against collection of stats. But it's the closest I've been able to find.

Perhaps you are misremembering?

[1] https://news.ycombinator.com/item?id=23735866

Another extremely unsatisfying possibility is that those threads could have been flagged as flame wars and de-indexed. I'm not entirely sure what happened, but I do distinctly remember reading those posts and I'm willing to stand by that.

I'm surprised to see your comment down-voted. I think that you precisely summarized the problem. Applying population statistics to individuals is problematic.

However to expand on that I think it is primarily a problem with point statistics. If a Bayesian philosophy was followed, it's about looking at the plausibility of the entire space and realizing an individual can fall anywhere with a non-zero plausibility (or even in some cases an individual can even fall into an area of 0 plausibility given that it is a model and not reality).

Applying statistics to individuals is not automatically bad. You may be interested if "your group" has inflated risks for certain cancers, for example.

but surely you want to make sure every PoC is individually compensated for the systemic racism they had to endure? How is that different?

I am not opposed to scrutinizing decision making. But that should happen at the human level, not by scrubbing information from the data or output of algorithms.

Some bias is also necessary for survival. When I am walking home at night, I am more likely to change the side of the street when a group of men is approaching me, than when a group of women is approaching me. You can tell me that is discrimination all you want, but I value my life higher than your ideals of political correctness.

Reply on news.ycombinator.com