White House Report on Big Data Discrimination

The White House has released a report on big-data discrimination. From the blog post:

Using case studies on credit lending, employment, higher education, and criminal justice, the report we are releasing today illustrates how big data techniques can be used to detect bias and prevent discrimination. It also demonstrates the risks involved, particularly how technologies can deliberately or inadvertently perpetuate, exacerbate, or mask discrimination.

The purpose of the report is not to offer remedies to the issues it raises, but rather to identify these issues and prompt conversation, research­ -- and action­ -- among technologists, academics, policy makers, and citizens, alike.

The report includes a number of recommendations for advancing work in this nascent field of data and ethics. These include investing in research, broadening and diversifying technical leadership, cross-training, and expanded literacy on data discrimination, bolstering accountability, and creating standards for use within both the government and the private sector. It also calls on computer and data science programs and professionals to promote fairness and opportunity as part of an overall commitment to the responsible and ethical use of data.

Posted on May 6, 2016 at 6:12 AM • 16 Comments

Comments

Recommendation 1: 6 p.s.i. Lat N 38 53 51.75, Lon W 77 2 11.55May 6, 2016 9:14 AM

US government Juche maintains that human rights (and increasingly constitutional rights) equal nondiscrimination. The government takes this line so they can implement any repressive measure as long as it's applied across the board. You have no other rights, they think, only the right to be pushed around exactly like everyone else.

The government will harp on this report in treaty bodies and charter bodies to evade review of government surveillance and repression of association, information, and expression. State department bureaucrats will wave it around May 13, when the Committee to End Racial Discrimination reviews US compliance with the binding commitments of the CERD. International human rights experts will make fools of them again, but the US delegation's public disgrace will be suppressed in statist media, so the subject population can continue to take their regime seriously.

Criminal justice case studies, nyuk nyuk. When the outside world asks why cops shoot black people with impunity, the US government tries to fix it by making sure cops shoot everybody with impunity. That's the purpose of this bullshit government report.

Duane Ramsdell "Dewey" ClarridgeMay 6, 2016 10:07 AM

A clever marketing attempt to sell mass surveillance as a tool for social justice while these collection techniques are used to mass murder innocent civilians and terrorize entire communities over in the Middle East.

This is how the media whitewashes war crimes. Hey, "Toxic Sludge is Good for you!"

Anyway, I'm sure the White House appreciates the shout out, Bruce. Perhaps it will provide a well known security celebrity with a Presidential photo op some day in the future. There's good money to be made in selling out.

DanielMay 6, 2016 10:50 AM

A clever marketing attempt to sell mass surveillance as a tool for social justice while these collection techniques are used to mass murder innocent civilians and terrorize entire communities over in the Middle East.

This.

I won't dispute that big data has its socially beneficial uses. What you will never convince me of, as things stand today, is that on balance those socially beneficial uses outweigh the overwhelming negative deleterious effects.

AndrewMay 6, 2016 11:39 AM

Wether we want or not, big data algorithms exists and they are working well. They simply cannot be ignored or their usage forbidden.
I think the report is ok, rising an alarm against their bias and asking for transparency and auditing.

Clive RobinsonMay 6, 2016 1:40 PM

The trouble with looking for discrimination, is you will always find it. There is "no mister average" for good reason, things always have a distribution, pick the wrong distribution for what you are measuring and you will see discrimination...

We have seen this in the slightly whaco branch of medical science called "nutrition". In general it is considered unethical to perform longterm possibly dangerous experiments on people, therefor they use "observation by self reporting" or analysis of other findings.

As some readers will be aware various diseases that appear in later life have been claimed to be on poor nutritional choices. One such was the Fat / Carb argument about which was most harmful.

Historical study suggests mankind is best suited by a low carb diet, not low fat. That is for all but the last few centuries refined carbs were very much the exception in human ingestion and only at certain times of year. Fat on the other hand along with protein formed the larger part of the calorific input from the "Hunter-Gatherer" life mankind had lived since before homosapiens became recognisable as a distinct species.

However, following the heart attack and stroke of a US President, certain individuals gained prominance pushing "low-fat, high-carb", despite other evidence this became the orthodoxy. Even the evidence that supposadly supported low-fat high-carb has since been found to have been incorectly analyzed. Likewise many of the studies once it had become the orthodoxy twisted themselves into statistical knots to show this, whilst opposing evidence simoly did not get published. It is now nearly half a century later that the opposing evidence is being published and gaining traction, and re-examination of past data showing invalid conclusions were reached likewise being published. However for a billion or so humans the harm has been done, their early death warrants signed by their doctors as part of the orthodoxy...

I realy don't expect this "discrimination" detection to work, and even if it does, any correction can only occure by "positive discrimination" which when all is said and done just further discrimination.

Further it's been found that "positive discrimination" when implemented, actualy does not achive what it sets out to do. Filling quotas automaticaly reduces the size of the pool of available candidates, which means on balance you are going to eliminate the larger section of candidates and thus probably the best candidates with them. The reason this happens is the quotas are trying to resolve the problem at the wrong end of the education/experience pipeline. You need to address the discrimination issues before the child is born, and that means social change over several --three minimum-- generations, there just are not any "quick fixes" to discrimination.

EhMay 6, 2016 4:06 PM

I like how our government TOTALLY ABANDONING the concept of all humans having basic human rights... is a complete non-issue... but oh, you dare to pay a woman or a black person less than a white man, and BOOM, big problems! We gotta have our priorities straight!

Discrimination MattersMay 6, 2016 6:13 PM

@Clive Robinson

I realy don't expect this "discrimination" detection to work, and even if it does, any correction can only occure by "positive discrimination" which when all is said and done just further discrimination.

Let's parse "just further discrimination". You clearly didn't mean 'just' in terms of 'justice', but it rather fits. Discrimination is f'ing awesome. Discrimination is how we survive past infancy. In all your attempts at insightful prose, you seem to have missed that foundation.

all caps trollMay 6, 2016 6:22 PM

@Eh

-1 obvious troll. Your all caps sentiment seems more apropo historically vis a vis slavery and Susan B Anthony et al. Unfortunately your obvious troll is probably still effective at negatively effecting the course of human history. Life is rough.

666May 6, 2016 6:28 PM

The report includes a number of recommendations for advancing work in this nascent field of data and ethics.

Sounds like a group of people that didn't watch star trek reruns in their youth.

the comment afterMay 6, 2016 9:31 PM

@DR"D"C

This is how the media whitewashes war crimes. Hey, "Toxic Sludge is Good for you!"

Indeed. Look everyone, chocolate rations are going up, doubleplusgood!

EhMay 6, 2016 10:39 PM

@all caps troll

Emphatically pointing out that we're doing far far worse things than discrimination and nobody cares, is trolling? Then let the trolling continue!

EhMay 6, 2016 10:44 PM

Oh, sorry... I should have said "Emphatically pointing out that we're doing FAR FAR WORSE things..." There, fixed it for ya :)

all caps trollMay 6, 2016 11:12 PM

@Eh

Sorry, but you clearly had gone beyond emphatically pointing out, into way stupid land. Which is how one drones a more correct argument to bits in a free speech forum when their own position is fatally logically flawed.

Slime Mold with MustardMay 7, 2016 10:50 AM

@Clive Robinson

You certainly hit my sore points. As I wrote on the Squid thread two weeks ago, science is very broken . Regrettably, the articles start with the "soft" sciences, where one would expect them. They do go on to medical research and even physics. But they barely scratch the surface.

The most popular manipulation nowadays is to selects data sets that support the (orthodox) conclusion, and ignore those that do not, even when the data sets have been discredited (although their discredit was not widely advertised - too many careers on the line).

A good example of this is Graybill's stripbark Bristlecone Pine series. A dendrochronology frequently cited as a "proxy" in climate studies. The idea being that; warm years should yield wider rings, cooler years, narrower. In 2007
Linah Ababneh published results contradicting this assumption. Since she had weather data available for more recent years, she was able to determine that ring width was determined exclusively by rainfall. But the Graybill series continues to be cited by researchers prominent, even famous, in the climate studies community (Michael Mann). There are other sheninanigans and especially "adjustment" fraud. I am trying to keep this short.

I have to fire people: Sometimes from firms they have worked for in excess of 20 years. (OK, I write the reports, conduct the interviews, use forensic methods or hire experts for physical evidence: It has been many years since anyone - administrative judges included - have questioned my conclusions). I start with the negative hypothesis . That is: What I suspect is not true . I take this to heart rather than as a silly opening paragraph. I am not merely open to other ideas. I actively search for alternate explanations.

RE: Discrimination

In the US, and I am told, the UK, young men of African ancestry face peer pressure against both education and meaningful employment. There are, of course exceptions. When I have to do hiring and or training, I often notice among young American black men an expectation of failure . I do not see this in many African American women, and almost never see it in those born in Africa. If you walk through the offices of America's cities, you will see three times as many black women as men. Some scholars have claimed that this is because the dominant white culture feels less threatened by women. I say we will never know unless we close this gapping cultural chasm. My firm employs black men in senior positions. With a single exception, they are all from Africa. I doubt their friends see their efforts at work as them "trying a be a white mother f***er".

The government's new tool, despite their protestation, will first be used as a basis of lawsuits, either bought by NGOs, the Department of Justice, Equal Employment Opportunity Commission, or Housing and Urban Development, and then eventually legislation. I have had to prepare the response to discrimination suits (always filed in bad faith). It never takes less than 200 hours of my time, and a similar amount from the company attorney. We have never paid a penny. The process is the punishment .

The AbyssMay 7, 2016 6:40 PM

@Slime ... Mustard

In the US, and I am told, the UK, young men of African ancestry face peer pressure against both education and meaningful employment. There are, of course exceptions. When I have to do hiring and or training, I often notice among young American black men an expectation of failure . I do not see this in many African American women, and almost never see it in those born in Africa.

A couple things this brings to mind- First, one can certainly understand the expectation of failure in the context of-

'We knew we couldn’t make it illegal to be either against the war or black, but by getting the public to associate the hippies with marijuana and blacks with heroin, and then criminalizing both heavily, we could disrupt those communities.'

http://www.thecannabist.co/2016/03/23/john-ehrlichman-war-on-drugs-richard-nixon-blacks-hippies/50631/

Second, I've had a hypothesis for awhile that Obama's lack(?) of direct U.S. slave ancestry, i.e. closer to 'born in africa' (lol), is responsible for his political viabiity. I.e. that perhaps there are enough voters that share your sentiment, and have less sense of threat from, or prejudice against that type of darker skinned person.

Leave a comment

Allowed HTML: <a href="URL"> • <em> <cite> <i> • <strong> <b> • <sub> <sup> • <ul> <ol> <li> • <blockquote> <pre>

Photo of Bruce Schneier by Per Ervland.

Schneier on Security is a personal website. Opinions expressed are not necessarily those of IBM Resilient.