Don't cut off data, add more!! Yes men crash more than women, but why? Given enough information about the individual, the algorithm will stop making guesses based on gender.
What "guesses"? Men are objectively more likely to crash. At the end of the day you are still working with probabilities, that's what insurance is about. Adding more data could help you make better estimates about likeliness of that particular instance of human crashing cars, but it cannot erase the underlying difference that objectively exists: men are more likely to get into traffic trouble.
Insurance used to group people by likelihood of them causing crash, besides gender, there is a huge dependency on: 1) age of the driver 2) model of the car 3) area (!!!) where car will be driven 4) experience of the driver
All of that is going to stay. We just happen to have started forbidding grouping by gender, so that Jane can get cheaper health insurance and have John pay more. But Jane is objectively both less likely to need doctors help AND more likely to visit a doctor nevertheless. John is also more likely to crash his car. Asking them to pay more for respective insurance was NOT discriminatory.
In the UK women used to get cheaper car insurance. That turned out to be illegal under gender equality rules. Insurance companies had to stop using gender as a risk factor. Women's car insurance went up, men's went down, in some cases a lot for younger drivers.
Not only UK, but EU and it was a side effect of introducing unisex tariffs in insurance in general. They were actually concerned with women having to pay more for health insurance, car insurance (tiny in comparison) is a by product.
There was no bias, however, men were objectively rarely going to see doctors, and women were objectively more risk averse when driving.
New firefighters have to undergo some tests before being accepted, both physical and academic. Those tests are designed to measure their likely performance at the job.
Same for almost anything. Skin colour rarely matters, and given enough more direct data on factors that do matter, skin colour will have no predictive value, so the algorithm will ignore it.
Exactly the opposite will happen. E.g.let's pick example that won't brush POC people wrong as it disadvantages men, car insurance.
You let AI know the gender - it will figure men are more likely to get into crash. If you stop feeding gender to AI, it will figure people named John are more likely to be causing trouble, than people named Julie. You will have to play a long cat and mouse game cutting off information sources for the AI to a point of it becoming useless.
Just think about the whole "discrimination" issue. Black and Latinos are poorer than White/Jewish/Asian americans. Hence AI, created to optimize the strategy, figured it makes more sense to target the latter with ads.
I mean, look, he had to "hold iphone" so that earbuds are in his ears (for some reason). Now, had they been of the wireless, sometimes exploding, type? No such problem!
Thank you, Tim Cook, for this amazing innovation! Wireless earbuds invention of Apple is second only to the Apple's invention of the internet!
Planes are expensive, airlines are naturally interested in making as many flights per day as possible.
But one could not expect too much from human beings entering the plane, it just takes time and I suspect it is boarding that takes longer than planned.
What if targeting is based on ROI of showing ads to particular groups completely handled by AI? Even if you forbid AI to use race or gender as a factor, it would figure that people called "Jane" are more interested in certain types of ads than people called "John". That people called "Mike" tend to have different behavior than those, called Jose.
I think we need to differenciate between human being projecting stereotypes to optimize ads (that's what "targeted" ads are) vs AI that is literally driven by ads effect doing it.
Targeted advertising is either legal or not. If it is, you'll always get "discrimination" mentioned in OP, which actually isn't discrimination at all.
An angry mob of several thousand people (at a company that employs about 100k) shut down AI Ethics meeting. But let's pretend "it's because 'mob was right'", or in other words #nothinghappened shall we?
The sad truth is that now Boeing is just as shitty as Airbus
Now I need to know when did Airbus:
1) Design a system which can decide to nosedive the plan 2) Make it depend on a single fucking sensor 3) Have guy who worked for airbus before, but is now owrking for the government, claim that sensor was supposed to fail "only" in 1 out of 100'000 cases hence it was OK to have "terminate this plane" feature depend only on that single sensor 4) After letting all that shit into production and actually killing people pretend nothing is wrong and do fucking nothing
I won't even dive into how the fuck is deadly piece of crap feature like that accompanied by only a fucking PDF being sent to pilots to read.
PS Oh, and no more "authorities faked data" BS please.
I think it's equally important to recognize that both NASA and the astronauts are putting the mission first.
Do they? This part sounded a bit different to me:
" McClain had thought she would be able to work in a large-size torso, but after her spacewalk last Friday, she wore a medium-size torso and learned that it fit her better. "
Facebook is lucky, as it is not a European company. It can just laugh and walk away. Companies in Europe must cause a blackout themselves or be blacked out.
Your naivety is adoring. Facebook will have to comply, as any other company doing business in EU.
But for that to happen local governments would need to adopt the laws.
If you really want to know about "reality", it is Scandinavian countries, Sweden in particular, that has actually looked into "male underrepresented" jobs. They actually spent money on programs encouraging men. Programs had a very short lived effect (of about 2 years).
No fucks are given in many countries. And, I recall when someone wanted to encourage boys to become veterinars (for practical reasons, the field, which is now dominated by women, needs people who could deal with big animals like cows, when most women stick with dogs/cats and smaller cuties), it was shut down, by... you guess it which group crying out loud about oppression and inequality.
Heck, even an attempt ot have a dude that would look into dude's mental health issues, was shut down, cause, equality.
There is. Getting equality for fathers, such as equal access to parental leave, is all part of an effort to get men to participate equally in child rearing. That in turn...
A fun fact: UK women group protesting against women and men retiring at the same age (they wanted to keep it old ways, when women would retire when 5 years younger) had "for equality" in their name.
because people are willing to trade a bit of cash for a better work life.
We see those people all over the place: 1) Certain people spend 30% less time commuting to job 2) Certain people are 20 times less likely to die at work 3) Certain people are doing fraction of overtime other people are doing
Guess which gender those certain people are? Plays very well with "gender gap" mythology, debunked decades ago, dead, but still alive, like zombies.
No, it's rather "not true feminists" argument. The theory was that choice gaps are mostly caused by some sort of oppression. Hence, once one progressed as society, logically, males/females differences would be ironed out.
Yet it's exactly the opposite of what had happened in Scandinavia.
But instead of accepting the obvious fact, of theory being, let me put it mildly, somewhat wrong, let's deny the very fact, shell we...
I guess the scientists trying to build practical fusion reactors are just continuing their victimhood narrative since they can'y actually be certain what the final design will be.
Success condition of fusion reaction is pretty damn clear, mind you: producing significantly more energy, than consuming.
The piece you commented on literally asked for "success condition".
Slashdot Mirror
Don't cut off data, add more!! Yes men crash more than women, but why? Given enough information about the individual, the algorithm will stop making guesses based on gender.
What "guesses"? Men are objectively more likely to crash.
At the end of the day you are still working with probabilities, that's what insurance is about.
Adding more data could help you make better estimates about likeliness of that particular instance of human crashing cars, but it cannot erase the underlying difference that objectively exists: men are more likely to get into traffic trouble.
Insurance used to group people by likelihood of them causing crash, besides gender, there is a huge dependency on:
1) age of the driver
2) model of the car
3) area (!!!) where car will be driven
4) experience of the driver
All of that is going to stay. We just happen to have started forbidding grouping by gender, so that Jane can get cheaper health insurance and have John pay more. But Jane is objectively both less likely to need doctors help AND more likely to visit a doctor nevertheless. John is also more likely to crash his car. Asking them to pay more for respective insurance was NOT discriminatory.
In the UK women used to get cheaper car insurance. That turned out to be illegal under gender equality rules. Insurance companies had to stop using gender as a risk factor. Women's car insurance went up, men's went down, in some cases a lot for younger drivers.
Not only UK, but EU and it was a side effect of introducing unisex tariffs in insurance in general.
They were actually concerned with women having to pay more for health insurance, car insurance (tiny in comparison) is a by product.
There was no bias, however, men were objectively rarely going to see doctors, and women were objectively more risk averse when driving.
New firefighters have to undergo some tests before being accepted, both physical and academic. Those tests are designed to measure their likely performance at the job.
And easier tests are created specifically for female firefighters, not to run into embarassing Women to become NY firefighter despite failing crucial fitness test
And your point was?
Same for almost anything. Skin colour rarely matters, and given enough more direct data on factors that do matter, skin colour will have no predictive value, so the algorithm will ignore it.
Exactly the opposite will happen. .let's pick example that won't brush POC people wrong as it disadvantages men, car insurance.
E.g
You let AI know the gender - it will figure men are more likely to get into crash.
If you stop feeding gender to AI, it will figure people named John are more likely to be causing trouble, than people named Julie.
You will have to play a long cat and mouse game cutting off information sources for the AI to a point of it becoming useless.
Just think about the whole "discrimination" issue.
Black and Latinos are poorer than White/Jewish/Asian americans.
Hence AI, created to optimize the strategy, figured it makes more sense to target the latter with ads.
WWII apparently didn't teach us anything.
Well, we didn't have wars of that scale since then, so perhaps it did.
I mean:
Google's message here is that moving to the cloud isn't just about data. It can also be about code, and the developers who write it.
What the hell is going on here, had there been some major "enjoy cannabis for free" event, perhaps?
I mean, look, he had to "hold iphone" so that earbuds are in his ears (for some reason).
Now, had they been of the wireless, sometimes exploding, type? No such problem!
Thank you, Tim Cook, for this amazing innovation! Wireless earbuds invention of Apple is second only to the Apple's invention of the internet!
Planes are expensive, airlines are naturally interested in making as many flights per day as possible.
But one could not expect too much from human beings entering the plane, it just takes time and I suspect it is boarding that takes longer than planned.
They wanted to know if Facebook was using their targeting algorithm where it should not be used.
Could you elaborate how a generic targeting algorithm can "discriminate".
What if targeting is based on ROI of showing ads to particular groups completely handled by AI?
Even if you forbid AI to use race or gender as a factor, it would figure that people called "Jane" are more interested in certain types of ads than people called "John".
That people called "Mike" tend to have different behavior than those, called Jose.
I think we need to differenciate between human being projecting stereotypes to optimize ads (that's what "targeted" ads are) vs AI that is literally driven by ads effect doing it.
Targeted advertising is either legal or not. If it is, you'll always get "discrimination" mentioned in OP, which actually isn't discrimination at all.
I'd prefer being spied by USA rather than China, thank you very much.
An angry mob of several thousand people (at a company that employs about 100k) shut down AI Ethics meeting.
But let's pretend "it's because 'mob was right'", or in other words #nothinghappened shall we?
t....okay maybe Barbara Streisand, but surely nobody else! ...
Maybe Barbara Streisand what?
The sad truth is that now Boeing is just as shitty as Airbus
Now I need to know when did Airbus:
1) Design a system which can decide to nosedive the plan
2) Make it depend on a single fucking sensor
3) Have guy who worked for airbus before, but is now owrking for the government, claim that sensor was supposed to fail "only" in 1 out of 100'000 cases hence it was OK to have "terminate this plane" feature depend only on that single sensor
4) After letting all that shit into production and actually killing people pretend nothing is wrong and do fucking nothing
I won't even dive into how the fuck is deadly piece of crap feature like that accompanied by only a fucking PDF being sent to pilots to read.
PS
Oh, and no more "authorities faked data" BS please.
ATTEMPTED TO RAPE A CLASSMATE
FFS, Ford wasn't his classmate, her accusation was messy at best and have changed over time on top of it.
I think it's equally important to recognize that both NASA and the astronauts are putting the mission first.
Do they? This part sounded a bit different to me:
" McClain had thought she would be able to work in a large-size torso, but after her spacewalk last Friday, she wore a medium-size torso and learned that it fit her better. "
Facebook is lucky, as it is not a European company. It can just laugh and walk away. Companies in Europe must cause a blackout themselves or be blacked out.
Your naivety is adoring.
Facebook will have to comply, as any other company doing business in EU.
But for that to happen local governments would need to adopt the laws.
Not only did it touch as single chart of the long list of facts entioned by Strumia, it still didn't escape hilarious bits:
"...I hence asked one of our PhD students, Tobias Mistele, to plot the same quantity..."
Confirmation bias away please.
If you really want to know about "reality", it is Scandinavian countries, Sweden in particular, that has actually looked into "male underrepresented" jobs.
They actually spent money on programs encouraging men.
Programs had a very short lived effect (of about 2 years).
No fucks are given in many countries. And, I recall when someone wanted to encourage boys to become veterinars (for practical reasons, the field, which is now dominated by women, needs people who could deal with big animals like cows, when most women stick with dogs/cats and smaller cuties), it was shut down, by... you guess it which group crying out loud about oppression and inequality.
Heck, even an attempt ot have a dude that would look into dude's mental health issues, was shut down, cause, equality.
There is. Getting equality for fathers, such as equal access to parental leave, is all part of an effort to get men to participate equally in child rearing. That in turn ...
Demonstrates hypocrisy of the said "equality" organizations, blocking shared parenting bills left and right.
A fun fact: UK women group protesting against women and men retiring at the same age (they wanted to keep it old ways, when women would retire when 5 years younger) had "for equality" in their name.
Ironic.
or about men falling behind in higher education
In US, boys are behind at all stages. And for decades.
because people are willing to trade a bit of cash for a better work life.
We see those people all over the place:
1) Certain people spend 30% less time commuting to job
2) Certain people are 20 times less likely to die at work
3) Certain people are doing fraction of overtime other people are doing
Guess which gender those certain people are?
Plays very well with "gender gap" mythology, debunked decades ago, dead, but still alive, like zombies.
it's a straw feminist argument
No, it's rather "not true feminists" argument.
The theory was that choice gaps are mostly caused by some sort of oppression.
Hence, once one progressed as society, logically, males/females differences would be ironed out.
Yet it's exactly the opposite of what had happened in Scandinavia.
But instead of accepting the obvious fact, of theory being, let me put it mildly, somewhat wrong, let's deny the very fact, shell we...
I guess the scientists trying to build practical fusion reactors are just continuing their victimhood narrative since they can'y actually be certain what the final design will be.
Success condition of fusion reaction is pretty damn clear, mind you: producing significantly more energy, than consuming.
The piece you commented on literally asked for "success condition".