Racism In Online Ad Targeting
An anonymous reader writes "Most of us are familiar with advertisements in online web searching, and by now we've grown accustomed to scrolling past the 'sponsored' results to get to the real responses to our query. And we know the ads are context-sensitive; for example, searching for our favorite Federation Starship will bring up ads for a similarly-named car-rental agency. But now a Harvard University professor has found a more disturbing trend in those contextual ads: racism. 'Sweeney says she has evidence that black identifying names are up to 25 per cent more likely to be served with an arrest-related ad. "There is discrimination in delivery of these ads," she concludes. Sweeney gathered this evidence by collecting over 2000 names that were suggestive of race. For example, first names such as Trevon, Lakisha and Darnell suggest the owner is black while names like Laurie, Brendan and Katie suggest the owner is white. She then entered these plus surnames into Google.com and Reuters.com and examined the ads they returned. Most names generated ads for public records. However, black-identifying names turned out to be much more likely than white-identifying names to generate ads that including the word "arrest" (60 per cent versus 48 per cent).'"
not an effect. Making use of stats is not racism. Racism is denying the fact that many blacks in the US have been disadvantaged and largely as a result are more likely to commit crimes and get arrested. Pretending the numbers don't exist is horrible.
But does anyone know what the most popular names associated with arrest rates happen to be? Maybe there's actual correlation there?
And I'm not saying that's correct either, since there is likely strong racial bias in many arrests made by officers themselves...
It's not racism if the ads were being generated purely on statistical models. It may be a sad reflection of the state of African-Americans, but unless you can show how people have manipulated the process for some reason, it's merely statistics.
Targeted ads are by definition discriminating.
Ascalante: Your bride is over 3,000 years old.
Kull: She told me she was 19!
isn't it racist to give them to your kids?
bickerdyke
Kit: The letter K appears in this script 1,456 times. That's perfectly divisible by 3.
Freddy: So what? So what you saying?
Kit: What am I saying? KKK appears in this script 486 times!
"The Greens lynched a hacker in Chicago. Last month, but I think the body's still hanging from the old Water Tower."
However, I have a keen proficiency when it comes to grasping the obvious. Black people make better running backs than white people. Asian's named Nguyen are more likely than white's named Cletus to get academic scholarships. At a Bar, an Hispanic male in his 20's is a much better percentage bet for party favors than a 50-something white woman. Should I be offended that Google singles me out for testosterone supplements?
Happiness in intelligent people is the rarest thing I know.
Ernest Hemingway
It's mustn't be easy being Green.
Circumcision is child abuse.
What's next? Serving up maxi-pad ads to female names?
I'm calling my congressperson.
Uh, no. Race exists in all species, it is a subdivision of that species, just like species is a subdivision of a genus. In dogs we call them breeds. Where the boundaries lie between races is somewhat arbitrary. Between these groups, physical and biological differences are the rule, not the exception. If it helps you to call them "ethnicities", go right ahead, because those differences are smaller but still exist.
This doesn't mean an individual is defined by his or her race, only that they tend toward different means than those of other races. Nor does it excuse discrimination under the law, or treating members of some races poorly simply because they're a member of that race.
This is scientific, biological reality. Even if all the races died except for one, in 50,000 years there would be multiple races again, provided different populations were allowed to separate entirely for a while.
I don't supposed she checked out the names of all the folks arrested. She might find that there is evidence there that suggests Google is merely following the numbers.
Now, she could argue that law enforcement is discriminatory. But that would be even more support for Google's efforts being evidence based and not a result of bias. It doesn't matter if you were arrested because of what you did or because of your name or color. You still need to make bail and get a lawyer.
When Fascism comes to America, it will call itself Anti-Fascism, and tell you to give up your guns.
I know pathetically stupid, loser asian people. I know highly inteligent and educated black people. I know a lot of people who most definitely defy the stereotypes which we are all very aware of. But there was a time when I would have defended the practices being described. I will not defend them. It's rather disgusting if this is the truth.
But all that said, stereotypes exist because of an overwhelming preponderance of anecdotal evidence. And in the end, advertisers want to get their ads in front of the people who would be the most interested in seeing them, So in "bad neighborhoods" we see billboards for bondsmen and lawyers among other things. And I have even seen this online when I am in areas with wifi in bad neighborhoods. While disgusting, it would not be practiced if it weren't also effective. So we shouldn't be asking "why are advertisers responding to statistical evidence?" We should be asking how can we make positive changes that result in a change in statistics!
Because if we are seeing to have a practice which is largely based on statistical evidence, we are merely seeking to institutionalize denial of the truth. I'm just not inclined to deny a problem when solving it would seem more appropriate.
I can't imagine how anyone with a name like that would have a personal bias in interpreting these results. No sirree.