Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

The Suspicion Machine

wired.com

Post reply on HN
Print view View original

21–30 of 32 posts

Re: The Suspicion Machine

#21

Post by tagyro »

0xDEAFBEAD wrote: post #19

The algorithm described in this article seems very bad. But I would argue that ML risk scores can, in principle, be better than human judgment. Humans seem more subject to bias than algorithms are. Algorithms only look at data, but humans are additionally vulnerable to stereotypes and prejudices from society. Furthermore, using an algorithm gives voters an opportunity to have a debate regarding how best to approach a…

> Humans seem more subject to bias than algorithms are.

One might think that, but algorithms are built by humans, so they (algorithms) automatically have the same biases as the humans that built them.

Re: The Suspicion Machine

#22

Post by nonrandomstring »

0wis wrote: post #20

I am not sure the data model is the problem here. I feel like the journalist tried really hard to make a case against scoring (which I do not like either), but overlooked the fact that the whole system in which its embedded is bad. The case should not be on the technology but on its usage. It’s already looking like a bad journalism piece in the first part : ” Being flagged for investigation can ruin someone’s life, a…

> problem was the processes around the algorithm not its existence... > the problem is not the algorithm, its the investigators.

Indeed. Going deeper, the 'problem' is a social/cultural belief that doing X at scale, using a computer, is somehow more ethical than having a bunch of people do the same immoral thing. Computer automation and algorithms become a moral justification (or at least a Hail Mary) for immoral acts.

There is at once a diffusion of responsibility, a causal disconnection of consequential acts, a reassignment of responsibility, and - given that we bow down to machines as our masters rather than our tools - a change from choice to a belief in the "inevitability" of unstoppable processes.

Together these make us no longer question whether:

1) Computers are more reliable, consistent and accurate than people

2) Computers are fairer/just

3) Computers are more comprehensive/inclusive or selective/prejudicial

4) Computers are actually economically more effective

Of course, this has been going on since the 1960s at least and was part of "systems analysis for automation". I think we have regressed. Whereas it was commonplace to sceptically question technology in the 60s and 70s, today we start with the assumption that it must be "better" and then have to figure out how maybe it's not.

Re: The Suspicion Machine

#23

Post by PH95VuimJjqBqy »

Biologist123 wrote: post #11
safety1st wrote: post #8

Earlier quoted context omitted.

This is like equating dick pic spam with sexual abuse, yes the first one is problematic, but no it is not equal to the second, in fact it's a disservice to victims of genuine abuse to equate them. I can get money from almost anybody who's willing to give it to me for any reason, the government's influence and visibility into this are limited. I can earn money in some other currency/jurisdiction and convert it, too. A…

Your analogy is a bit of an odd one, as unsolicited dick pics are undoubtedly a form of sexual abuse.

Are you implying their point about equivalency is incorrect?

The other poster is not saying random dick pic's are ok, they're saying there's a gradation to the depravity.

It's like calling a 25 y/o man who sleeps with 17 y/o girl a pedophile. You're using the same word that you would if the girl was 10, but those are not nearly the same acts. One is inappropriate, the other is heinous.

this "guilt inflation" just makes it so many of us stop trusting words.

When I hear the word pedophile, I shouldn't have to ask clarifying questions to find out if the speaker is trying to manipulate my feelings or not. Likewise, if someone uses the phrase "sexual assault", I shouldn't have to ask clarifying questions to determine if it was some jackass sending unsolicited dick pics online or something more harmful (that may or may not also involve sending unsolicited dick pics).

Because, just like the pedophile example, a dick pic from a rando online and a dick pic from your boss or coworker are two wholly different acts.

Re: The Suspicion Machine

#24

Post by denton-scratch »

There's something schizophrenic about constructing a system for choosing who to investigate, while simultaneously trying to avoid discrimination.

The entire purpose of the chooser system is to discriminate between people; they want to investigate only those people likely to be cheating. If they really want to avoid discrimination, then they chould be choosing who to investigate using a straw-poll.

They have laws against certain kinds of discrimination, e.g. on the basis of race or gender. If those facts are used as input to the chooser, then race- and gender-discrimination is inevitable. There's not usually any protection against discrimination for e.g. being short, or having red hair, or speaking with a regional accent; I have no idea how such characteristics are correllated with cheating on welfare claims.

Re: The Suspicion Machine

#25

Post by Jensson »

nonrandomstring wrote: post #22
0wis wrote: post #20

I am not sure the data model is the problem here. I feel like the journalist tried really hard to make a case against scoring (which I do not like either), but overlooked the fact that the whole system in which its embedded is bad. The case should not be on the technology but on its usage. It’s already looking like a bad journalism piece in the first part : ” Being flagged for investigation can ruin someone’s life, a…

> problem was the processes around the algorithm not its existence... > the problem is not the algorithm, its the investigators. Indeed. Going deeper, the 'problem' is a social/cultural belief that doing X at scale, using a computer, is somehow more ethical than having a bunch of people do the same immoral thing. Computer automation and algorithms become a moral justification (or at least a Hail Mary) for immoral act…

The major difference is embarrassment. People don't like when other people see them in private, but they don't care if computers see them in private, so computerized surveillance is a lot more acceptable to people.

Re: The Suspicion Machine

#26

Post by nonrandomstring »

Jensson wrote: post #25
nonrandomstring wrote: post #22

Earlier quoted context omitted.

> problem was the processes around the algorithm not its existence... > the problem is not the algorithm, its the investigators. Indeed. Going deeper, the 'problem' is a social/cultural belief that doing X at scale, using a computer, is somehow more ethical than having a bunch of people do the same immoral thing. Computer automation and algorithms become a moral justification (or at least a Hail Mary) for immoral act…

The major difference is embarrassment. People don't like when other people see them in private, but they don't care if computers see them in private, so computerized surveillance is a lot more acceptable to people.

Good point. The benign neutrality supposition is a powerful factor I forgot. It was Weizenbaum who first wrote about this when noticing the relation between his secretary and the Eliza program.

Re: The Suspicion Machine

#27

Post by adammarples »

tagyro wrote: post #21
0xDEAFBEAD wrote: post #19

The algorithm described in this article seems very bad. But I would argue that ML risk scores can, in principle, be better than human judgment. Humans seem more subject to bias than algorithms are. Algorithms only look at data, but humans are additionally vulnerable to stereotypes and prejudices from society. Furthermore, using an algorithm gives voters an opportunity to have a debate regarding how best to approach a…

> Humans seem more subject to bias than algorithms are. One might think that, but algorithms are built by humans, so they (algorithms) automatically have the same biases as the humans that built them.

I think there's a case to be made that data contains bias, ie. previous arrests, encoding the bias of the past arresting officer, but the algorithms don't contain bias. Unless you could explain how logistic regression is biased.

Re: The Suspicion Machine

#28

Post by 0xDEAFBEAD »

tagyro wrote: post #21
0xDEAFBEAD wrote: post #19

The algorithm described in this article seems very bad. But I would argue that ML risk scores can, in principle, be better than human judgment. Humans seem more subject to bias than algorithms are. Algorithms only look at data, but humans are additionally vulnerable to stereotypes and prejudices from society. Furthermore, using an algorithm gives voters an opportunity to have a debate regarding how best to approach a…

> Humans seem more subject to bias than algorithms are. One might think that, but algorithms are built by humans, so they (algorithms) automatically have the same biases as the humans that built them.

That doesn't follow.

If I'm a chemist, and I write an algorithm to do something related to chemistry, that algorithm does not "automatically" know everything I know about chemistry.

Bias works the same way.

Re: The Suspicion Machine

#29

Post by CommitSyn »

thedailymail wrote: post #12
CommitSyn wrote: post #5

Earlier quoted context omitted.

I desperately want to read this article. Anyone else getting /constant/ cloudflare CAPTCHAS while trying to access archive.ph? Android with mobile Firefox or chrome, first with wifi then mobile then VPN. No matter what I can't access it and it won't stop showing the "are you a human?" check page, even after trying the CAPTCHA the first time it stops showing. Must be a cloudflare outage?

If you are using Cloudflare 1.1.1.1, try switching it off or using another DNS resolver. There are reported issues with archive.md blocking access from 1.1.1.1

Thanks I was unaware of that, but I use AdGuard and Quad9.

Re: The Suspicion Machine

#30

Post by CommitSyn »

15457345234 wrote: post #6
CommitSyn wrote: post #5

Earlier quoted context omitted.

I desperately want to read this article. Anyone else getting /constant/ cloudflare CAPTCHAS while trying to access archive.ph? Android with mobile Firefox or chrome, first with wifi then mobile then VPN. No matter what I can't access it and it won't stop showing the "are you a human?" check page, even after trying the CAPTCHA the first time it stops showing. Must be a cloudflare outage?

You don't need to read the article, it seems like you understand the problem :)

From the title alone I certainly understand the issue, but this looks like news on said subject I understand, which I would like to read.
Post reply on HN