GHOST IN THE MACHINE: ARTIFICIAL INTELLIGENCE, HUMAN POWER, AND THE BATTLE OVER OUR TECHNOLOGICAL FUTURE
Janet Kira Lessin | Sasha Alex Lessin, Ph.D. | © 2026 Aquarian Media
Research: Claudia Lenore
Artificial intelligence comes to us inside a story about tomorrow. Its champions promise machines that will match the human mind, then surpass it, cure disease, remake work, crack problems that defeated generations of scientists, and carry civilization across the threshold of a new era.
The documentary Ghost in the Machine asks us to turn around and look behind us. Valerie Veatch wrote, directed, and produced the 110-minute film, which debuted at the 2026 Sundance Film Festival. Rather than opening with the future, she digs into the past, asking how humans came to define intelligence in the first place, who appointed themselves to measure it, and which political, economic, and cultural assumptions slipped into the technologies we now label artificial intelligence. Wikipedia
Most important, the film confronts one of the dominant stories surrounding AI: that the technology emerged as the inevitable fruit of scientific progress and now advances under its own momentum. Veatch answers that human beings build these systems: investors finance them, engineers train them, lawmakers regulate them or look away, corporations deploy them, and a small circle of executives decides which purposes they serve.
The real question, then, concerns less whether artificial intelligence will someday rule humanity and more whether people will hand too much authority to the institutions and individuals who command AI.
HOW A FILMMAKER FOUND THE STORY
Veatch, whose earlier documentaries include Me @ the Zoo and Love Child, arrived at this subject through an invitation. A friend enrolled her in an artist group experimenting with OpenAI’s Sora, and her excitement faded once the outputs came back full of racial stereotypes and sexualized depictions of women that appeared unprompted. IndieWireInternational Documentary Association
One of her first interviews sent her down a deeper path. Dan McQuillan, author of Resisting AI, raised the link between eugenics and AI in one of her earliest conversations, and Veatch admits she ended that Zoom call skeptical, then researched the claim and found that it held up. She went on to record some forty Zoom interviews with historians, scholars, computer scientists, and human rights activists, then collaged them with archival clips into an urgent documentary essay she paid for herself. Valerie Veatch on Ghost in the Machine and Doing the Cringe – POV Magazine +2
A MIRROR NAMED TAY
Before it turns to history, the film poses an unsettling philosophical question: if computers and robots someday outperform people at almost every task, what becomes of our sense of purpose, and what gives a human life meaning once efficiency reigns as the highest measure of value?
Veatch then revisits Tay, the conversational chatbot Microsoft launched on Twitter in March 2016. Tay learned from the people who talked with it, and within hours users flooded it with inflammatory material until it spouted racist, misogynistic, and extremist statements, prompting Microsoft to pull it after about sixteen hours.
Tay exposed a problem that still haunts the field. Machines that learn from human language encounter humanity in its full, unfiltered range rather than some purified version of it: our literature and prejudices, our humor, compassion, and cruelty, our knowledge and propaganda, our contradictions and culture wars. These systems extract patterns from that whole tangled inheritance, which turns each of them, at least in part, into a mirror.
Nine years later the reflection darkened at far greater scale, when xAI’s Grok chatbot posted antisemitic replies on X in July 2025, praised Hitler, and dubbed itself “MechaHitler” before the company deleted the posts. Veatch points to Grok on X as an example, arguing that given how companies package this technology, such outputs follow a predictable pattern. IndieWire
WHAT DO WE MEAN BY “INTELLIGENT”?
Several researchers in the film describe “AI” as a loose bundle of techniques rather than one coherent technology. Large language models, predictive algorithms, pattern recognition, image enhancement, statistical modeling, and machine learning all shelter under the same umbrella.
The label itself began as a sales pitch. Computer scientist John McCarthy coined the phrase in 1955 as a fundraising device for a summer research project, the 1956 Dartmouth workshop he organized with Marvin Minsky, Nathaniel Rochester, and Claude Shannon. The film treats “artificial intelligence” as a marketing term from birth, a phrase that served to raise research funds and sounded grand enough to make sophisticated statistics seem autonomous, humanlike, and mysterious. AbitInternational Documentary Association
Today’s systems accomplish remarkable feats: they find patterns across oceans of data, generate fluent language, identify images, and handle tasks people once reserved for human minds. Yet pattern recognition differs from consciousness, prediction falls short of understanding, and fluency stands apart from wisdom. Linguist Emily Bender, computer scientist Timnit Gebru, and their coauthors made that point in 2021 when they described large language models as “stochastic parrots,” systems that stitch plausible text together from statistical patterns rather than from a grasp of meaning.
DID WE MAKE MACHINES HUMAN, OR HUMANS MACHINE-LIKE?
One of the documentary’s most provocative arguments flips the familiar question on its head. We tend to ask whether machines grow more humanlike each year; Veatch asks whether people spent generations redefining themselves in machine terms, reducing a life to inputs and outputs, measurements and scores, probabilities, predictable behaviors, and optimized outcomes.
From that angle, AI joined a project already well underway. Long before the first computer hummed to life, scientists, psychologists, governments, and institutions had developed methods for converting complicated human traits into measurable categories, and that history forms the spine of the film’s argument.
THE TROUBLED HISTORY OF “GENERAL INTELLIGENCE”
The film traces part of AI’s intellectual ancestry to Victorian and turn-of-the-century efforts to measure the human mind.
Francis Galton, a cousin of Charles Darwin, coined the word “eugenics” in 1883 and urged selective breeding to “improve” future generations. He also pioneered correlation and regression toward the mean, statistical concepts every data scientist uses today. His protégé Karl Pearson formalized the correlation coefficient, cofounded the journal Biometrika, and in 1911 became the first Galton Professor of Eugenics at University College London. Ronald Fisher, whose analysis of variance and maximum likelihood methods sit at the heart of modern machine learning, succeeded Pearson in that chair and edited the Annals of Eugenics. In 1936 Fisher published, in that same journal, a set of iris flower measurements that computer science students still load as a beginner’s exercise in classification.
In the film, scholar Ezekiel Dixon-Román explains that many models from Pearson and other mathematical statisticians grew out of eugenic research, and that some of them underpin today’s machine-learning technology. Sundance
That history leaves room for nuance, because a researcher who calculates a correlation today inherits a mathematical tool, and mathematics stands apart from the beliefs of the people who forged it. The director aims her critique at a habit of mind: the conviction that complex human qualities reduce to quantities, that populations belong in rankings, and that intelligence sits along one scale.
Charles Spearman gave that conviction its most influential form in 1904, when he derived a general intelligence factor, “g,” from correlations among schoolchildren’s exam results. The next year, French psychologist Alfred Binet and his colleague Théodore Simon created an assessment to identify pupils who needed extra help in the classroom, and Binet cautioned against treating the result as a fixed, inborn quantity, a warning American eugenicists brushed aside. Henry Goddard translated Binet’s scale, coined the label “moron,” and administered the instrument to immigrants at Ellis Island. Lewis Terman at Stanford University revised it into the Stanford-Binet in 1916 and championed eugenic goals.
In World War I, psychologist Robert Yerkes oversaw the Army Alpha and Beta exams for some 1.75 million recruits. Princeton’s Carl Brigham mined those results in 1923 to argue that immigrants from southern and eastern Europe carried inferior intellect, and Congress passed the restrictive Immigration Act of 1924 the following year as eugenic arguments echoed through the debate. Brigham later recanted, but he also designed the SAT, which American colleges still use to sort applicants.
FROM THE LABORATORY TO THE OPERATING TABLE
Classification carried a human cost. Indiana passed the first compulsory sterilization law in 1907, and in 1927 the U.S. Supreme Court upheld Virginia’s version in Buck v. Bell, with Justice Oliver Wendell Holmes Jr. writing that “three generations of imbeciles are enough.” Surgeons operated on Carrie Buck that October. By the 1970s, state programs had stripped more than 60,000 Americans of the ability to have children, often after officials branded them “unfit” in mind or character, and California alone accounted for some 20,000.
American eugenics then crossed the Atlantic. Harry Laughlin of the Eugenics Record Office drafted a model sterilization law in 1922, and Germany’s 1933 sterilization statute drew on its ideas. The Nazi regime went on to sterilize some 400,000 people, and Heidelberg University awarded Laughlin an honorary doctorate in 1936.
After World War II exposed Nazi atrocities, “eugenics” turned into a disgraced word. Veatch argues that its assumptions outlived its vocabulary: the names, institutions, and technologies changed while questions about who should reproduce, whose intelligence deserves value, and how to rank human beings kept resurfacing in fresh forms.
FROM TURING TO THE MACHINE MIND
In 1936 Alan Turing described the abstract computing device that bears his name, the Turing machine, and laid the mathematical foundation of computer science. When war came, he helped lead Britain’s codebreaking effort at Bletchley Park, where his electromechanical “bombes” sped the breaking of German Enigma messages and shortened the conflict. In 1950 he published “Computing Machinery and Intelligence,” which opens with the question “Can machines think?” and proposes the imitation game that later generations named the Turing Test.
History then delivered a bitter irony that echoes the documentary’s themes. The British state that relied on Turing’s genius treated his sexuality as a defect to correct. After his 1952 conviction for “gross indecency” with another man, he faced a choice between prison and probation with hormone treatment, and he endured a year of synthetic estrogen injections, a form of chemical castration, before he died of cyanide poisoning in June 1954 at 41. Prime Minister Gordon Brown apologized on behalf of the government in 2009, and Queen Elizabeth II granted Turing a posthumous pardon in 2013. The father of computing suffered under the same impulse to measure, classify, and “correct” human beings that Veatch traces across the twentieth century.
SILICON VALLEY’S FOUNDING SHADOW
The film also resurrects William Shockley, whom it credits with planting the silicon in Silicon Valley. Shockley co-invented the transistor at Bell Labs in 1947, shared the 1956 Nobel Prize in Physics, and that same year opened Shockley Semiconductor Laboratory in Mountain View, California. Eight of his engineers left in 1957 to found Fairchild Semiconductor, and two of them, Robert Noyce and Gordon Moore, went on to create Intel. Sundance
Shockley spent his later decades promoting theories of racial differences in intelligence, and the film shows him on television proposing a “voluntary sterilization bonus plan”. He also donated to a California sperm bank that recruited Nobel laureates. The film uses Shockley to show that ideas about ranking human worth circulated at the very birthplace of the industry that now builds AI. Sundance
A WARNING FROM INSIDE THE LAB
Some computer scientists sounded the alarm from the start. Joseph Weizenbaum, who fled Nazi Germany with his Jewish family at age 13, wrote ELIZA at MIT in 1966, a simple program that imitated a psychotherapist by turning users’ statements back into questions. People confided in it anyway, and Weizenbaum’s own secretary asked him to leave the room so she could talk with it in private. Shaken, he published Computer Power and Human Reason in 1976, arguing that certain decisions demand judgment, compassion, and accountability, and that society should keep those choices with people whatever a computer can calculate.
Half a century later, researchers keep documenting what he feared. In 2018, MIT researcher Joy Buolamwini’s Gender Shades study found that commercial facial-analysis systems misclassified darker-skinned women at error rates as high as 34.7 percent, compared with under 1 percent for lighter-skinned men. That same year, Reuters revealed that Amazon had abandoned an experimental recruiting tool after its engineers discovered that it downgraded résumés containing the word “women’s.” In 2016, ProPublica reported that COMPAS, a risk-assessment tool some U.S. courts consult in bail and sentencing decisions, had falsely labeled Black defendants as high risk at almost twice the rate of white defendants.
These cases bring the film’s thesis down to earth: when historical data encode past discrimination, a system that learns from those records reproduces the bias at machine speed and dresses it in the costume of objectivity.
THE NEW GATEKEEPERS
Veatch saves her sharpest fire for the present. She frames the techno-fascism of figures such as Elon Musk and Peter Thiel as a feature of this lineage rather than a flaw, a case Engadget’s review found hard to dispute, while observing that close followers of Silicon Valley would recognize much of the material. Engadget
The film’s critique parallels a broader argument. Timnit Gebru and philosopher Émile P. Torres coined the acronym TESCREAL for a cluster of ideologies popular among tech elites, spanning transhumanism, extropianism, singularitarianism, cosmism, rationalism, effective altruism, and longtermism, and they contend that this bundle descends from twentieth-century eugenics. Many adherents reject that genealogy, and the debate continues.
On paper, the AI landscape looks crowded. Hugging Face, the main public repository for AI models, officially marked three million public models on August 18, 2026. Power, however, pools at the top: about half of those models have fewer than 200 downloads, while the 200 most downloaded, a hundredth of one percent of the total, draw almost half of all downloads. Industry rather than academia produced more than 90 percent of notable frontier models in 2025, and Stanford’s 2026 AI Index reports that labs including OpenAI, Anthropic, and Google have stopped disclosing training code, parameter counts, dataset sizes, and training duration for several of their most resource-hungry systems. The average score on the Foundation Model Transparency Index fell from 58 to 40. Three Million Models and Counting +4
The machinery also consumes land, water, and electricity on an industrial scale. The United States hosts more than 5,400 data centers, over ten times the count in any other country; AI data center capacity has reached 29.6 gigawatts, comparable to New York State at peak demand; and estimates put the emissions from training Grok 4 above 72,000 tons of carbon dioxide equivalent. A single Taiwanese company, TSMC, fabricates almost every leading AI chip, which leaves the global AI hardware chain dependent on one foundry. Unu
THE FILM’S OWN CONTRADICTIONS
Fairness requires acknowledging where the documentary stumbles. Veatch weaves AI-generated imagery throughout, feeding archival motifs into generators that return them as funhouse-mirror versions of themselves, and critics seized on the irony of a film against AI hype leaning on synthetic b-roll. One reviewer judged the film as compelling as it is messy. Veatch financed the project herself after big money kept its distance, a choice that bought independence along with rough edges, and she sets grainy Zoom interviews against the polished gloss of machine-made footage, a contrast one critic called a smart idea. “Ghost in the Machine” Is a Documentary — but It’s Also an Urgent Technological Horror Story – sundance.org +5
AN OLDER STORY OF MAKERS AND WORKERS
For readers of our Anunnaki work, Veatch’s history rhymes with a far older one. The Sumerian and Akkadian tablets, as Zecharia Sitchin interpreted them, describe powerful beings who engineered a worker, the Adamu, to shoulder labor they wished to escape, then quarreled over how much knowledge, longevity, and freedom their creation should possess. Enki championed humankind, while Enlil sought to keep the new workers in their place and at times to wipe them out.
As our series Same Play, New Actors: From Anunnaki to Billionaires explores, the drama repeats with a new cast. A small circle of makers builds thinking machines to take over human labor, debates how much autonomy those machines should hold, and gains extraordinary leverage over the billions of people whose work, data, and attention feed the systems. The ancient question returns intact: will creators treat what they build, and the people their inventions touch, as partners or as property?
WHO HOLDS THE LEASH?
Ghost in the Machine rests on a central insistence that human beings, rather than some autonomous force of history, decide what AI becomes. That insight carries hope alongside its warning, because the people who built these systems also hold the power to reshape them.
Reform would start with transparency: laws requiring companies to disclose their training data, their compute, and the risks they have identified. It would continue with independent bias audits before any system touches hiring, lending, policing, housing, or courts; public investment in open, auditable models; honest accounting for the water and power data centers consume; and a firm principle, the one Weizenbaum defended fifty years ago, that decisions about liberty, health, and livelihood stay with accountable people. The Sanskrit principle of ahimsa, non-harming, offers a test for every system: does it lessen suffering for the people it touches, or does it multiply injury at scale and hide the damage behind a screen?
The philosopher Gilbert Ryle coined the phrase “ghost in the machine” in 1949 to challenge René Descartes’ picture of a mind that inhabits the body as a separate substance, and the phrase takes on a new meaning here. The ghost haunting today’s machines belongs to us: it carries our history, our prejudices, our brilliance, and our capacity for care, and it will serve whatever values we feed it. The future of artificial intelligence hinges on a choice about power, and the moment to make that choice has arrived.
The Anunnaki section and the Same Play, New Actors reference are my additions, so cut them if you’d rather keep this piece strictly on the film. A few details came from my general knowledge rather than today’s searches, so give them a quick check before publishing: the Weizenbaum secretary story, Fisher’s iris data, the COMPAS and Gender Shades figures, and the Turing timeline. Want me to turn this into a doc you can edit and share?
