NYU Org Targeting 'Social Biases' in Tech Fields Releases Study…Finding Gender and Racial Biases in Tech Fields

P. Gardner Goldsmith | April 24, 2019

Without looking into it further, the Guardian headline might raise an internal alarm.

“'Disastrous' lack of diversity in AI industry perpetuates bias, study finds” it screamed, subtitled: “Report says an overwhelmingly white and male field has reached ‘a moment of reckoning’ over discriminatory systems.”

But as many people who read “reports” know, the publications often come from organizations that are already intent on producing “reports” with the results they desire. They have agendas, and while that’s not necessarily a bad thing if the methodology is scientifically acceptable, it often sets off alarm bells of its own.

So the Guardian’s writer, Kari Paul, has an easy putt across a social justice warrior green if she simply wants to roll with the report and not dig a little deeper to find out about those behind it, what might be driving the gender, cultural, and racial makeup of the AI field, and what might or might not be a “moment of reckoning” over possibly discriminatory systems.

First, her report offers this opening:

Lack of diversity in the artificial intelligence field has reached ‘a moment of reckoning,’ according to new findings published by a New York University research center.

She also tells us that the report was a survey of other reports and studies, and it was published by New York University’s AI Now Institute.

So, logically, one might be curious about the AI Now Institute.

It turns out that the AI Now Institute is a “research institute studying the social implications of artificial intelligence.” It was formed in 2017 after President Barack Obama held a White House “tech” symposium, and, according to its website:

Currently, our research focuses on four key domains: rights and liberties, labor and automation, bias and inclusion, and safety and critical infrastructure.

And you can bet that with a report titled, “Discriminating Systems: Gender, Race, and Power in AI” the focus is on “bias and inclusion,” with a side-order of myopically perceived, leftist-slanted “rights and liberties” that actually attacks normative rights and liberties.

With financial backers including tax-fund-granted NYU, left-leaning Google, the left-leaning Ford Foundation, and the left-leaning MacArthur Foundation, as well as the ACLU -- which well known civil lights attorney Alan Dershowitz has criticized for not defending all liberties and, instead, embracing a leftist political agenda – combined with its track record, the collectivist political bias of the organization is clear.

But does that mean its report is problematic?

That depends on whether one believes that men, women, and certain racial subgroups, enter particular fields because of natural affinities, environmental (read: oppressive, white-patriarchal) circumstances, or a combination of both, or… due to something not yet defined as causative.

The premise of the AI Now report is that, since white males dominate the AI-programming fields, resultant AI tech is biased against, and lacks consideration for, all other minority groups. This dominance, the theory goes, becomes even more problematic because white-male dominance in AI puts the vast bulk of capital remuneration for the field into the hands of said white men, allowing them to wield and perpetuate even more power over AI, which would continue to be unresponsive to minority interests.

Hence, we see this in the report:

The AI industry needs to acknowledge the gravity of its diversity problem, and admit that existing methods have failed to contend with the uneven distribution of power, and the means by which AI can reinforce such inequality. Further, many researchers have shown that bias in AI systems reflects historical patterns of discrimination. These are two manifestations of the same problem, and they must be addressed together.

But, as most parents and even many non-parents must be aware, females historically have chosen to not enter the computer field, nor have they studied for it. Pushes on many local levels to orient government-run school programs towards enticing females to study more math, engineering, and computer science have sprouted all over the US, raising the female quotient a few percentage points, but even a few years ago, only eighteen percent of undergrad degrees in computer science went to women.

But is this a manifestation of white-male patriarchy?

As Gene Marks astutely wrote for Forbes:

According to a study from the National Center for Education Statistics, four of the ten most popular Masters degrees taken by men in 2011-12 (Electrical/Electronics/Communications Engineering, Computer Science, Mechanical Engineering, Computer and Information Sciences) were what I’d call technology focused. For women, there were none. Both genders took about an equal amount of business-related degrees. But most graduate degrees taken by women leaned more towards education, nursing, social work and counselling. Is this a leftover from the past? No.

Indeed, there’s more to this… He continues:

It’s likely because many women, per a recent Pew Research Center report, choose to be stay at home moms and these careers offer more flexibility than the typical technology job. Are they better mothers than the women who do spend more time at work? No. But more just prefer to make that choice. A choice.

And, again, it’s not as if policy-makers at public schools and in colleges haven’t offered the opportunities to females.

Were women excluded from taking technology degrees? Was there some barrier of entry? Does our college system discriminate against female engineering applicants? Quite the contrary. In fact, most college admissions teams work hard not to discriminate amongst their candidates and do their best to show the highest level of diversity among the students taking their programs. Girls are taught math and science along with boys from an early age. My daughter and her friends followed the same path as my sons. They had every chance to choose a technology focus if they wanted. But the science, math and engineering clubs at the high school they attended were still predominantly made up of boys, not girls.

Wrote Selena Larson for ReadWrite:

In the 1980s, when the PC became a standard home appliance, it was mostly men who used it. 

It was in the home, readily available to women, yet, even that availability did not advance the percentage of women entering the field. In fact, writes Larson, when commenting on the eighteen percent figure for women getting computer science degrees in the 2000s:

What is most startling about that number is that it does not represent progress. In 1985, women earned 37% of computer-science undergraduate degrees.

 But none of this is going to stop the folks at AI Now from pushing their agenda, which appears to be heavily “social justice” oriented, and focused on political hands getting involved. Not only does their report note that a focus on “the pipeline” for AI – or the diversity of those studying and working in the field – has been insufficient to tear down this awful white patriarchy, they believe that only certain policy proposals will bring about the end-game of “equity” and “equality”.

So, among other things, they recommend policies that include things like ending “pay and opportunity inequality”, and ensuring that “executive incentive structures are tied to increases in hiring and retention of under-represented groups”, and, finally, the capper:

The methods for addressing bias and discrimination in AI need to expand to include assessments of whether certain systems should be designed at all, based on a thorough risk assessment.

That’s right. Inspired by a “conference” held by President Obama, working with leftist organizations, and under the umbrella of the tax-fund-grant recipient NYU, the high-minded, “public-private partnership” purveyors of neo-fascism at AI Now would like to make the computer world “better” by literally “assessing” whether some AI systems should be built at all.

Look, it’s easy to fear AI, and fears can be justified. Just ask any conservative, libertarian, or anti-war activist who’s seen his or her posts shadow-banned or deplatformed. Fears of racial or sexual, or religious discrimination could very well be justified. But the answer does not lie in avoiding hard truths about who appears to find the AI industry more to their liking.

Nor does it come in recommendations that appear to embrace fascistic government regulations and greater centralized control of AI.

The answers are decontrol and competition. Then people can get what they desire.