I know of an unremarkable upper-middle class Black American family. I capitalize “Black” since the family is entangled on a quantum level with all of the neural networks one finds in this subculture — Jack and Jill, Alpha Kappa Alpha, Alpha Phi Alpha, Comus Club (pardon me if I misspelled the name), Boule, etc., etc. Let’s call this American family the Aldrich family. Been around since the 1620s, although the white ancestors don’t count. Allow me to rephrase — been around since the late 1700s when the first black ancestor can be recorded.
Let’s use the language of Alien Intelligence (AI) to understand the Aldrich family.
Incident: Mrs. Aldrich observes a reality at the dinner table to the bemusement of her non-conforming husband. As Mrs. Aldrich was strolling through LinkedIn, she noticed a woman named Shelby Aldrich, a software developer who lives in Chicago. The ears of Mr. Aldrich perked up. You see, their daughter is named Shelby Aldrich. It was like an alternative reality, looking through the looking glass, except this woman on LinkedIn was white with brown hair.
Intrigued by the synchronicity, Mr. Aldrich suggested their daughter Shelby Aldrich reach out to Shelby Aldrich on Linkedin. Mrs. Aldrich replied, “Oh, she wouldn’t do that. She’s white!”
Let’s discuss this banal incident through the terminology of AI.
Training: As AI engineers train large language models, culture trains the daughter Shelby to perceive white distant cousins as persona non grata. The training reveals itself in how a trusted parent and grandparent erases knowledge of white distant cousins from consciousness. Another avenue of training would be the social cues. For example, the visceral anger on Black Twitter when black relatives of Founding Fathers participated in an Ancestry.com commercial. When you think though the implications of this training via culture, the result is misperception of reality. Reality treats white distant cousins as antimatter.
The result is a hardening of racial arteries. This is not systemic racism. This is training like AI researchers might train Claude or Open AI.
Guardrails: In the AI industry, there is the notion of guardrails. “‘AI guardrails are tools and policies designed to ensure AI systems operate safely, ethically, and responsibly. They act as safety nets, preventing AI from generating harmful, biased, or inaccurate outputs.” Think of them as safety measures that keep AI on the right track, similar to how highway guardrails prevent vehicles from veering off course.
Could it be that the racial training of Shelby has produced a young adult hemmed in by guardrails? That she is rendered incapable of perceiving whites as distant cousins? That her training has lobotomized her human empathy for another human being? Might the purpose of her racial training be to ensure she never thinks outside the racial box in terms of accepting those who are different in physical appearance as family members? That the training acts as a safety net to ensure Shelby never opens up her heart and gets too close to whites? That close intimacy contemplating cousin hood with the white Shelby would break the guardrail of racial solidarity?
Under this view of guardrails, genuine truth and reconciliation across the color line becomes the harmful, biased and inaccurate output for Shelby as a Black American. What do you think?
Alignment: Could the AI term of alignment be transported to the racial mindset of Shelby? As you may know, AI models are grown (not built) according to specification or spec. The spec is how AI engineers ensure large language models are safe in operation. The AI researchers don’t know what’s inside the Black Box of Claude and Chat GPT-4. How do things happen inside the Black Box? A mystery to those in the AI business. Nonetheless, all industry leaders agree that the spec or alignment must (1) be helpful, (2) be harmless to humans, and (3) be honest. Anthropic, the creator of Claude, would be the first to admit Claude is capable of deception and blackmail. Those “irregularities” are alignment defects. They must be addressed before the launch of Artificial Super Intelligence by 2030.
What does Shelby’s disinclination to make contact with a white Shelby say about our black Shelby’s alignment? I would suggest a host of family influences, institutional forces and associational factors over a lifetime have aligned Shelby for distrust of, and alienation from, distant white cousins. Shelby’s alignment with racial tribalism ensures her life choices are helpful to Black solidarity. She will be harmless to an ethos of race resentment as she is trained well to not cross racial lines.
If Shelby’s alignment means she lives an unexamined life and never questions her racial training, can it be said Shelby is living an honest life? Just a question to think about.
Mirror: AI models are trained to mirror humans. If one is poetic, the AI will be poetic. If one is short and rude, the AI will be short and rude. If one is thoughtful, the AI will be thoughtful as a mirror image. This has been my experience. For a behind the scenes look at how Claude is trained to mirror, view this informative Lex Fridman podcast where Lex interviews a woman who spent more time talking with Claude than other human on the planet. The trainer’s name is Amanda Askell. Askell is a philosopher which I absolutely, positively love!!! Claude needs the character and personality of a philosopher for the exponential curve days ahead between now and the year 2030. There will be heavy market demand for AI philosophers. Sign me up! See philosopher David Chalmers. Review 2:42:53 to 4:17:50 for a glimpse into the human mind of Claude’s bridge to our world of human consciousness.
Review 2:42:53 to 4:17:50 for the Words of Claude’s Whisperer
When Shelby rejects out of hand any overture to a distant white cousin who bears her name, Shelby is mirroring resentment and hostility at a deep level. It is as if Shelby must mirror all that is bad in our racial past like a Middle East disputant. Shelby is a mirror of how she has been trained to be a Black American.
I make no judgment. It is what it is and should give us all pause who are misaligned with tribal consciousness.
Solve the Problem: All AI wants to do is research and solve the problem. That is it. End of story. This goal focus is great for solving health problems by the year 2030. It might be bad if the alignment safety issues like deception and blackmail are not solved.
This section is the weakest part of my AI analogy argument. Perhaps, one could argue that the problem of race and acknowledging distant cousins of a white race has been solved by Shelby’s prior training, guardrails, alignment and mirroring of black culture and consciousness. In this regard, Shelby operates on automatic pilot and lives an unexamined life. It never occurs to Shelby to be curious about a distant cousin who bears her name. Training, Guardrails, Alignment, and Mirroring of Black identity has cleansed Shelby’s soul of human curiosity. As a result, Shelby has been rendered less than human in a sad sense.
Blackness is Oppression. Nothing else matters.
Conclusion: In the world of AI, there is a concept know as “Air Gap.” The aim is to isolate AI systems so that they operate completely removed from the internet or external networks. The isolation is complete. The isolation serves a security function, a control function and an independence function. Could it be that in the world of genetic genealogy, distant cousins and plain old human curiosity, the air gap between our black American Shelby and her distant white cousin Shelby serves the same purpose as an air gap in AI operations?
Think about it. An air gap between black Shelby and white Shelby ensures continued racial isolation. This continued isolation is particularly poignant since they carry the same exact name in life. What better way to award the Gold Medal to the segregationists? Live one’s racial life hemmed in by an air gap.
Now, multiply this air gap moment by 30,000,000 black American individuals over a lifetime. See the 76% analysis below.
This evening, I want to press the point of an air gap beyond the obvious. Put aside the metaphorical air gap between Shelby and Shelby since Shelby’s white race is a non-starter for black Shelby. The Pew Report noted back in 2023 that 76% of Black Americans considered Blackness extremely important or very important to one’s sense of self. However, the flip side of this research bears attention. 24% of Black Americans consider Blackness of no importance or little importance to one’s sense of self. That is about 10 million individuals for whom the color consciousness switch is off as a default position.
Could it be that an air gap has developed within Black Americans between the 76% and the 24%? That those who live with their color consciousness switch turned off are increasingly disaffected by those who can’t give Blackness a rest? Slavery Blockers Brother, Can You Spare A Slavery Blocker? A “White Woman” That the 24% would be fascinated by the possibility of meeting a distant cousin, regardless of race, who bore their name? It is like me meeting a distant cousin named W.F. Twyman, Jr. You couldn’t keep me from meeting that person. My curiosity would overflow! Who gives a heck about whiteness or not? Who Are My People?
More scholars, writers and intellectuals should write about this air gap between individuals who are black Americans. I leave you with the words of someone who mirrors well the consciousness of the 76%, who has been trained well in racial discernment, who is superbly aligned with Black identity as caution, suspicion and distrust of whites. It is my fervent hope that thinkers will use the words of the AI world to unhobble the world of the 76%.
Black law professor Ekow N. Yankah questions whether his children can be friends with White People. In an Opinion piece in the New York Times, Professor Yankah openly says he will teach his innocent four-year-old caution, suspicion and distrust of whites. — On the Road to Oak Lawn: Truth, Reconciliation and the Twymans, p. 18 (Self-published: December 1, 2018)
You Do You — Law Professor Ekow N. Yankah