I wasn't going to post this because it's not directly related to RPGs, but since the thread has been moved to Pundit's forum, it may help provide some context.
Below is a section on the brain's immediate response to race from Sapolsky's
Behave, which is an attempt to synthesize current (2017) understanding of how the human brain functions, from biology through psychology and culture. It draws on all major relevant fields, like neurology, endocrinology, psychology, anthropology, sociology, and so on. This isn't a book about race, it's just comes up a few times as part of how humans think.
Sapolsky is a neurobiologist and primatologist, so he's coming at this from the perspective of biology first and social sciences second. But he does think biology and psychology and culture are inexorably intertwined. He's not deterministic; while the brain has inherent tendencies, he presents a lot of evidence that the specifics can be shaped by environmental factors. He doesn't seem to be particularly woke (I'd guess he's probably a Bill Maher style traditional US liberal).
This section is on the immediate response to stimuli. Seconds or less, not centuries or even minutes. As I mentioned before, the immediate responses to race appear to be real, but the brain operates at multiple levels over multiple time frames. The quote is largely about the amygdala, which is inaccurate but fast. Slower and more socially complex responses can override this, like ones from the pre-frontal cortex, not to mention things like culture.
There's more information on how humans response to race in other sections. Much of the response to race seems to be part of the human tendency to group people into Us and Them; in other words it's a specific manifestation of a much broader classification tool. It's also not necessarily aimed at race in specific; Sapolsky points out that our hunter-gatherer ancestors were unlikely to travel far enough to run into people of other races, so any evolutionary pressure to identify races would be weak or non-existent. But the broader tendency to group people into Us/Them categories is deeply embedded in humans, appearing almost at birth, and is omnipresent in how we think. He is highly skeptical of the broad claimed utility of implicit bias tests. He does reference the Clark doll study, but it's a single paragraph in an 800 page book, mostly focused on the self-hating aspect.
The numbers are footnotes, which I can copy if desired.
A hugely unsettling sensory cue concerns race.7 Our brains are incredibly attuned to skin color. Flash a face for less than a tenth of a second (one hundred milliseconds), so short a time that people aren’t even sure they’ve seen something. Have them guess the race of the pictured face, and there’s a better-than-even chance of accuracy. We may claim to judge someone by the content of their character rather than by the color of their skin. But our brains sure as hell note the color, real fast.
By one hundred milliseconds, brain function already differs in two depressing ways, depending on the race of the face (as shown with neuroimaging). First, in a widely replicated finding, the amygdala activates. Moreover, the more racist someone is in an implicit test of race bias (stay tuned), the more activation there is.8
Similarly, repeatedly show subjects a picture of a face accompanied by a shock; soon, seeing the face alone activates the amygdala.9 As shown by Elizabeth Phelps of NYU, such “fear conditioning” occurs faster for other-race than same-race faces. Amygdalae are prepared to learn to associate something bad with Them. Moreover, people judge neutral other-race faces as angrier than neutral same-race faces.
So if whites see a black face shown at a subliminal speed, the amygdala activates.10 But if the face is shown long enough for conscious processing, the anterior cingulate and the “cognitive” dlPFC then activate and inhibit the amygdala. It’s the frontal cortex exerting executive control over the deeper, darker amygdaloid response.
Second depressing finding: subliminal signaling of race also affects the fusiform face area, the cortical region that specializes in facial recognition.11 Damaging the fusiform, for example, selectively produces “face blindness” (aka prosopagnosia), an inability to recognize faces. Work by John Gabrieli at MIT demonstrates less fusiform activation for other-race faces, with the effect strongest in the most implicitly racist subjects. This isn’t about novelty—show a face with purple skin and the fusiform responds as if it’s same-race. The fusiform isn’t fooled—“That’s not an Other; it’s just a ‘normal’ Photoshopped face.”
In accord with that, white Americans remember white better than black faces; moreover, mixed-race faces are remembered better if described as being of a white rather than a black person. Remarkably, if mixed-race subjects are told they’ve been assigned to one of the two races for the study, they show less fusiform response to faces of the arbitrarily designated “other” race.12
Our attunement to race is shown in another way, too.13 Show a video of someone’s hand being poked with a needle, and subjects have an “isomorphic sensorimotor” response—hands tense in empathy. Among both whites and blacks, the response is blunted for other-race hands; the more the implicit racism, the more blunting. Similarly, among subjects of both races, there’s more activation of the (emotional) medial PFC when considering misfortune befalling a member of their own race than of another race.
This has major implications. In work by Joshua Correll at the University of Colorado, subjects were rapidly shown pictures of people holding either a gun or a cell phone and were told to shoot (only) gun toters. This is painfully reminiscent of the 1999 killing of Amadou Diallo. Diallo, a West African immigrant in New York, matched a description of a rapist. Four white officers questioned him, and when the unarmed Diallo started to pull out his wallet, they decided it was a gun and fired forty-one shots. The underlying neurobiology concerns “event-related potentials” (ERPs), which are stimulus-induced changes in electrical activity of the brain (as assessed by EEG—electroencephalography). Threatening faces produce a distinctive change (called the P200 component) in the ERP waveform in under two hundred milliseconds. Among white subjects, viewing someone black evokes a stronger P200 waveform than viewing someone white, regardless of whether the person is armed. Then, a few milliseconds later, a second, inhibitory waveform (the N200 component) appears, originating from the frontal cortex—“Let’s think a sec about what we’re seeing before we shoot.” Viewing a black individual evokes less of an N200 waveform than does seeing someone white. The greater the P200/N200 ratio (i.e., the greater the ratio of I’m-feeling-threatened to Hold-on-a-sec), the greater the likelihood of shooting an unarmed black individual. In another study subjects had to identify fragmented pictures of objects. Priming white subjects with subliminal views of black (but not white) faces made them better at detecting pictures of weapons (but not cameras or books).14
Finally, for the same criminal conviction, the more stereotypically African a black individual’s facial features, the longer the sentence.15 In contrast, juries view black (but not white) male defendants more favorably if they’re wearing big, clunky glasses; some defense attorneys even exploit this “nerd defense” by accessorizing their clients with fake glasses, and prosecuting attorneys ask whether those dorky glasses are real. In other words, when blind, impartial justice is supposedly being administered, jurors are unconsciously biased by racial stereotypes of someone’s face.
This is so depressing—are we hardwired to fear the face of someone of another race, to process their face less as a face, to feel less empathy? No. For starters, there’s tremendous individual variation—not everyone’s amygdala activates in response to an other-race face, and those exceptions are informative. Moreover, subtle manipulations rapidly change the amygdaloid response to the face of an Other. This will be covered in chapter 11.
So yes, babies show racial bias. This isn't hard-coded because different people have different responses, an initial reaction may be very different from a response that takes a longer period of time (the longer period may be less than a second), and it's not the same thing as racism.
The whistleblower is wrong in denying this, and so are the trainers, who are drawing unsupported conclusions.