The Major Flaw in AI… That No One Will Talk About

This morning, my partner Darlene and I were engaged in our Saturday morning ritual: sitting on our back porch, drinking coffee, and talking. Suddenly, she raised a topic that “lit up” my mind and made me sit bolt upright.

She said she feared that a fundamental flaw in AI might prove our undoing, as a species. But it was her line of reasoning that excited and worried me. As a woman, and a highly intelligent person who gets computers, she knows that AI is only as good or reliable as its programming — and the vast majority of programmers are men steeped in patriarchal or even misogynistic thinking. Not to mention, male arrogance and hubris.

This line of thinking inspired me because it explains a LOT about the developing, often antisocial impacts of AI. (Many tech bros and “broligarchs” actually envision a future in which humanity is replaced by sentient AI.) Darlene reminded me about a personal story she’d told me before — something rather shocking she experienced personally in her workplace.

“Techies” and programmers behaving badly

At the time, she worked as a supervisor in a high-tech services company that helped other professionals with their IT issues. All the programmers and tech trouble-shooters who worked in her department were young men, who often seemed quite arrogant. One day, she overheard several of them complaining about a doctor’s office that wasn’t very “tech savvy,” and they were making fun of the doctors and their staff. They mockingly labeled them “stupid” and “clueless.”

Darlene was horrified, both by their attitudes and the smug certainty behind them. She recognized a teachable moment and approached the group. Then, she calmly asked them, “Do any of you know how to perform a surgery or operation?” Of course, they all said they didn’t.

She then pointed out that their company’s clients were professionals who knew a LOT about their chosen field — just like the programmers. Next, she asked, “So, do any of you think you could DO a surgery, without extensive training?” Again, they all (sheepishly) said “No.”

Finally, she finished her “intervention” by reminding them how much we all depend on the expertise and knowledge of these experts — just as those experts depend on her employees’ training and knowledge. She reminded them that these are the exact people the group was ridiculing and portraying as “stupid” — and that their strong need for IT services didn’t make them “clueless.” She advised them to revise their attitudes and be less judgmental, more respectful, and more accurate in their assessments.

Apparently, they got it. She never heard them talk that way again.

Press enter or click to view image in full size

AI’s built-in “domination mindset”

This story is quite revealing. Not only does it nail the supreme arrogance of some techies and programmers, but it also — and even more importantly — points out the tech world’s crucial need for direct feedback/intervention, at times, as well as ongoing education in interpersonal dynamics and relating.

Yet in almost all high-tech circles and among AI developers, it’s clear that hubris, greed, a “macho” dominator mindset, and vastly over-inflated expectations rule the roost. Worse, there is NO countervailing narrative or any serious attempt to “humanize” the direction of AI.

There are, of course, many unserious, superficial claims that AI is “fair” and totally unbiased. And without a potent counter-narrative to “prove” that major biases are built into AI systems, the dominant players and arrogant “broligarchs” now get the final word.

Even the sort of “efficiency” touted by AI proponents is inherently biased, since it takes as a given that more and more computer-driven “efficiency” is automatically good for corporations and the overall economy — even if hundreds of thousands of jobs are eliminated.

This antiquated, super-capitalist, and wildly unfair version of “efficiency” is based on an endless growth, profit-is-everything model that benefits top management and shareholders while oppressing workers. This flawed model flows directly from an aggressive, greedy, domination-oriented mindset.

Meanwhile, gentler, more humane, and ecologically friendly concepts of efficiency are either sidelined or treated with derision and contempt. And the planet and workers suffer the dire consequences.

Current AI systems are reactionary, NOT revolutionary

Why is the built-in bias or one-sidedness of AI a huge problem? Well, for most people, it isn’t — and that’s the exact problem. The “power-over” dominator mindset is so enmeshed with our society and integrated into our thinking that few people ever challenge it or are fully aware of it.

Yet AI is always touted as an “extraordinary leap” in human comprehension and abilities. The hype surrounding AI practically proclaims it a panacea for mankind’s problems. Admittedly, it’s much faster than prior computing, for sure, and can perform many millions of computations and deliver complex results in seconds.

But, so what? Is it really “far better” if we can get a hundred million bits of bullshit and biased crap per second, instead of a mere million? Is a huge truckload of stinking, rotting manure really “better” than a cupful?

Darlene’s and my great fear is that the AI “revolution” is, in truth, just a vastly sped-up and intensified version of the same old, dysfunctional systems and reactionary thinking patterns. In fact, the AI “revolution” seems to be paralleling the intensifying social push, in the US and around the world, toward greater misogyny, racism, and vast wealth inequities.

This is not a great mystery. AI is moving in these directions and supporting our current dark (anti)social drift because AI’s foundations are tainted, imbalanced, and wrong-headed. AI is fundamentally unsound precisely because it’s built on a very limited, domination-oriented view of “reality.”

Perhaps AI only reflects and projects the worldview of the male half of humanity — primarily white men — while totally ignoring or dismissing the views and understanding of the other half (far more than half, really).

That’s a travesty. If AI is to “help humanity progress,” it would need to be rethought and thoroughly revamped. That would be truly revolutionary.

—

This post was previously published on Fourth Wave.

***

You Might Also Like These From The Good Men Project

You Said ‘Race’, but Are You Actually Talking About Race?

Understanding the Nonbinary: Are You Confusing Gender With Sex?

The Difference Between Compassion for Those With Disabilities & Ableism?

‘Masculinity’ Is Having an Identity Crisis