Eugenics Powers IQ and AI

Both the definition of “intelligence” and the tech industry are deeply entwined with white domination. Will white-supremacist AI be the result?

In another era of technological innovation and widening inequality, when engineering to revolutionize society was also in vogue, an eerily familiar cadre of mad tinkerers unleashed their vision for humanity to devastating effect. As anti-Black and anti-immigrant violence shook the US, these earlier “disruptors” popularized the notion of “intelligence”: the idea that all humans have innate and fixed abilities that can be accurately assessed, measured, and used to categorize people for a lifetime. Like the wizards of tech today, these disruptors were Very Smart People, their self-esteem and economic viability reinforced, often since “gifted” childhoods, by over-indexing on traits like verbal comprehension, perceptual reasoning, processing speed, and working memory. Unquestioningly devoted to their defiantly Eurocentric perception of the world, these white men implanted their reductive definition of human ability into the cores of the intelligence tests they created and spread through American society.

As culturally specific and scientifically invalid as their definition has been judged in the intervening century, this idea of “intelligence” continues to be bought and sold as the essence of human reason, the one that defines human potential and capability through the measurement of “IQ.” Today’s tech industry is the golden child of this “intelligence.” And as sweet as it is to possess the diagnosis and paycheck of a Very Smart Person, there is no denying that the Very Smart definition of “intelligence”—like the DNA of tech itself—is deeply intertwined with the white-ethnic domination championed throughout the sciences in the 19th and 20th centuries.

As we hurtle, ever faster, into a future shaped and slicked over by artificial intelligence, don’t take your eye off this weight that bears us ever back into the violent racial contradictions of our past and present. The weight is white, embedded in every supporting document of the institutions where white power and privilege reside. This weight distorts every turn, every code, every byte of the aggregated and calibrated store-bought “intelligence” that streams hourly through our fantasy of liberatory technology and self-cleaning robotic reality. The weight is carried in invisible backpacks by bots stocking our shelves and cleaning our hospitals, algorhythming our shopping lists and harvesting our foods, measuring out our flesh and blood in units of “value” and “capacity.” This weight demands—and guarantees—that our technologies, however clever, never realize their promises of shiny, efficient social equality.

This racist weight—putting its thumb on the scale of every facet of our technology—feels eternal. But it is actually new and malleable, created in a strange, recent, forgotten, and denied history. It is a racist history, yet one that we carry forward—silently—within one of society’s core concepts: an “intelligence” that fetishizes speed, efficiency, and innovation; a narrow, hierarchical “intelligence” that only pays lip service to the “soft skills” that might better cultivate equity—empathy, creativity, communication, collaboration, altruism. The idea that this “intelligence” (encompassing discrete capacities to sort, categorize, process, and remember verbal and visual information as accurately and quickly as possible) is a universal and essential human trait, then, goes without saying.

It is eugenics that secretly sits at the heart of both IQ and AI.

This skeletal rendition of “intelligence,” crucially, entitles and empowers Very Smart People to revolutionize society according to their own needs and whims, whether political, social, economic, or even emotional. Lost in this neatly reductive understanding is the degree to which this “intelligence” is not nearly as scientifically valid—nor as essential to the survival of humanity—as its proponents would like us to believe.

The study of “intelligence” emerged during a time like our own, another Saturn-Pluto synod. Then as now, global capitalistic pressures and conflicts were boiling, and white panic in the face of movements for racial justice and reintegration was creating chaos in the white American soul and polity. Like tech, the study of intelligence was birthed from a tricky, shape-shifting alliance between archetypical white heroes and villains: nerdy innovators, earnest “helpers,” ruthless social Darwinists angling for power and wealth, and various amalgams of the three. To a powerful segment of 20th-century (and 21st-century) psychologists and other scientists—and their institutional benefactors—the idea that every human has a comparable innate intelligence, which could be developed through effort and appropriate instruction, was anathema to these elites and their goals for American society.

The last century of European and American scientific study of intelligence is synonymous with the championing of the superiority of white intelligence. The project of using emerging statistical and psychometric methods to reinforce the racialized notion that all humans are not created equal became eugenics. Its influence over our 21st-century self-image of intelligence and humanity is profound but also hidden.

It is eugenics that secretly sits at the heart of not only IQ but AI. Unless there are radical changes, the next century will bring the championing of the superiority of artificial white intelligence, and the reification of its power.


Understanding Race with AI

By Bruce D. Haynes et al.

The father of Silicon Valley founder Frederick Terman was one of the most celebrated psychologists in American history, Lewis Terman. From his half-century perch at Stanford University, the elder Terman vigorously injected the idea of a measurable, scientifically validated, racialized “intelligence” into the American consciousness. Terman’s lifetime of writings and advocacy makes clear the racist underpinnings of his ideas. He was deeply invested in a fantasy of hierarchical intelligence in which “gifted,” mostly white children are groomed for leadership and influence, and everyone else is slotted into supporting roles in the industrial machine. Intelligence, as Terman conceived it in 1916, is narrow, fixed, hereditary, and variable across ethnic groups, and always higher among white and more affluent children. This notion of intelligence lives on robustly in the IQ tests and other standardized measures that psychologists, educators, and employers use today.

Despite the scientific community’s eventual rejection of overt racism—something that Terman himself never acknowledged or accounted for—Terman’s contributions continue to shape our popular and scientific understandings of human ability, more than half a century after his death. Terman apologists today focus on his devotion to a meritocracy built on measurable ability. But the most statistically reliable element of Terman’s work has been the reproduction of the old paradigm of social dominance. This paradigm is encoded in a categorical bias toward the mentally “strong” over the mentally “weak”: “white” over “black,” “native” over immigrant,” “rich” over “poor,” “male” over “female.”

Terman’s long tenure at Stanford, and his family’s ongoing intellectual influence in Silicon Valley, extend his legacy into new questions around artificial intelligence. And the continuing dominance of Terman’s “intelligence” forces us to ask how AI, as conceived today, could do anything but reinforce Terman’s false promises of a whites-on-top meritocracy—accompanied by just enough “diversity and inclusion” exceptions to prove the rule.

Follow the template that Terman and colleagues laid down: that human intelligence is universal, hierarchical, measurable. Next, posit that human intelligence is the superior intelligence. Should you follow this logic—and create artificial intelligence in the image of the human brain—then what we will eternally return to is an “intelligence” that privileges whiteness.

Social history is also family history. And the Terman family history is the story of how blind whiteness can be to itself, and to its unconscious devotion to social domination coated in the gilt of grandiose altruism.

Lewis Terman told himself and the world that he was using statistical methods to understand why some children are more “gifted” than others, and thus helping them to help make the world a better place. What Terman was actually searching for was a rationale for why he, as a white northern European man born into a family of modest means in the early 20th century, was given the keys to the kingdom. What he found was an explanation of how he, and people like him, think, and why they believe that their way of thinking makes their people superior to all others.

Armed with that information, Terman and his family transformed the world to their own benefit, but not everyone else’s. The Terman family arc crystalizes the many ways that tech and intelligence resonate on a wide historical frequency. But, consequently, their family history is also the key to understanding how eugenic ideas of intelligence lie in wait, ready to sabotage any attempt to undo tech’s inborn legacies of racialized bias and inequity.

In fetishizing “gifted” white children, the Terman family sought a “rational” explanation for their undemocratic urge to pull the ladder up behind them.

In the last century, Terman’s signature Stanford-Binet Intelligence Scales have been used for identifying learning disorders; justifying the limitation, marginalization, and even termination of thousands of lives though forced sterilization; educational and professional tracking; and restricting immigration from “undesirable” countries.

Yet Terman’s own family faced no such limitations. Indeed, Terman’s fetishization and defense of “gifted” white children was as much a personal origin story, rooted in the racially segregated social isolation he experienced as a nerdy kid in late-1800s rural Indiana, as scientific truth.

In one generation, Terman’s Scotch-Irish tenant-farmer family was unshackled and transformed, from “backward” (Terman’s word) to affluent and influential on a scale that endures today. Even as Lewis Terman and his students were limiting the opportunities of those they considered “dull” due to “the family stocks from which they come,” Terman’s own family was rocketing through the social and economic ranks of postwar white America. Their trajectory—and Terman’s “scientific” explanation of its basis in “giftedness”—mapped perfectly onto the shared desires of 20th-century white Americans. What they sought was a “rational” explanation for their undemocratic urge to pull the ladder up behind them and hoard resources and opportunities from immigrants more recently arrived than themselves.

“I know of nothing in my ancestry that would have led anyone to predict for me an intellectual career,” Terman later wrote. “A statistical study of my forebears would have suggested rather that I was destined to spend my life on a farm or as the manager of a small business, and that my education would probably stop with high school graduation or earlier.” Terman was one of 14 children; his father had attended school only a few months a year.

Terman’s own children and grandchildren had a very different fate. Terman’s son, Fred, studied electrical engineering at Stanford, became dean of the engineering school and then university provost, and paved the way for the creation of Silicon Valley. Terman’s grandson Lew followed in his father’s footsteps, receiving multiple engineering degrees at Stanford and spending four decades in research at IBM, eventually becoming president of the company’s Academy of Technology.

(Fred Terman publicly neither embraced nor rejected his father’s eugenicist views. Even so, his fellow Silicon Valley founder, Nobel laureate and Stanford professor William Shockley, passionately defended eugenics until his death in 1989. Shockley’s adherence to Terman’s eugenicist views had an ironic twist: he was tested as an elementary schooler by Terman’s researchers, but his scores were too low to qualify as “gifted.”)

Terman continues to enjoy the support of surprisingly prominent academic apologists; they defend Terman’s core theories, while maintaining that eugenicists were a product of “their time.” Despite such shameless support in the present day, there was robust contemporary resistance to Terman’s ideas in his own time.

I hate the impudence of a claim that in fifty minutes you can judge and classify a human being’s predestined fitness in life,” wrote journalist Walter Lippmann in 1922, debating Terman in the New Republic. “I hate the abuse of the scientific method which it involves. … I hate the sense of superiority which it creates and the sense of inferiority which it imposes.” Lippman’s passionate critique was taken up, in the decades that followed, by psychologists and educators, who were struck by the severe limitations of the eugenics-inspired understanding of intelligence. Unsurprisingly, conceptualizing intelligence in a way that privileges a narrow style of “gifted,” generally white and European information processing doesn’t do much to address the naturally occurring diversity of human learning styles. Nor does such a concept rectify the economic and racial inequities that are themselves byproducts of the biases underpinning a eugenic definition of intelligence.

Starting in the 1970s, a field of countertheories of intelligence bloomed, planted by superstar academic researchers like Howard Gardner, Robert Sternberg, and Daniel Goleman. There are now books and papers and theories galore that counter Terman’s narrow vision. These champion emotional intelligence; triarchic intelligence; linguistic, logical-mathematical, musical, bodily-kinesthetic, spatial, interpersonal, and intrapersonal intelligence. All these ideas bring us closer to a more holistic and complete understanding of human capacity.


Designing AI with Justice

By Sasha Costanza-Chock

And yet, in the most impactful ways, Terman and the eugenicists won the debate on intelligence. They are still winning.

Terman’s theories, and the culture of intelligence testing and measurement they inspired, dominate our practical understanding and application of human ability. And they do so even as the aspirational language of civil rights, multiculturalism, diversity, access, and equity permeates our social discourse.

The instruments and practices derived from Terman’s work—IQ tests, standardized tests (including the SAT), Gifted and Talented programs—continue to promote and solidify a hierarchical distribution of power. This hierarchy is grounded in a definition of human value that repetitively and predictably exalts and damns various segments of the population. We have cleansed the racial language from the study of intelligence—and eugenicists from the history of psychology and the sciences—but left the eugenic core unscathed in theory and practice.

The result is still eugenics, but without eugenicists. Just as sociologist Eduardo Bonilla-Silva explained that “racism without racists” still reproduces a racially unjust society, “eugenics without eugenicists” produces inequitable hierarchies of power. Until there is a full recognition, reckoning, and repair of the racist origins of intelligence theory—root and branch—any application of Terman’s theories of intelligence will inevitably achieve his desired result.


This article was commissioned by Mona Sloaneicon

Featured image: Untitled (detail) (2019). Photograph by Brett Jordan / Unsplash