Are algorithms shaping reality more than facts do nowadays?
Civic Witness Series: Questions from the Attic, is founded on a simple premise: the people living through democratic erosion have questions the ‘experts’ aren’t addressing; it is inspired by and in honor of Adia in the Attic. Those questions matter because they provide an opportunity for sharing and documenting what democratic erosion feels like to people living through it. We don’t pretend to have “The Answer” to these questions, but we hope to stimulate the conversation her questions have prompted.
We’re currently living in a time when arguments about the dangers or benefits of artificial intelligence (AI) are pervasive, with both sides often going to rhetorical extremes to make their points. Like any new technology, AI has positives and negatives. Not unlike the arguments I remember about television in the early fifties or, later, social media and mobile phones. At some point in their evolution, each of these was seen as either the best thing since sliced bread or the end of civilization as we know it.
Though not an expert, my involvement with computers dates to the mid-fifties; I built an analog computer from a kit I ordered through the mail when I was about eleven. Later, I was somewhat involved in the development and deployment of the field artillery digital fire direction computer in the mid-sixties, and finally, in the vanguard of installing the now-obsolete Wang word processing computers at our diplomatic posts in China in the early eighties, I am familiar with both the upside and downside of automated systems. I know, for instance, that algorithms, the set of rules designed to help systems solve problems or perform computations, can seem like magic to the uninitiated. And because they work quietly in the background, they can shape a person’s view of reality, notwithstanding the existence of data that proves something different.
This is not a new phenomenon.
Let’s start with a little reverse psychology. Consider the 1952 presidential election, when UNIVAC predicted an Eisenhower landslide that initially seemed too improbable to release. In this case, the machine was right. This was not a case of an algorithm spreading falsehood, but of people hesitating to trust a computer’s interpretation of evidence because it challenged their expectations.
The lesson, then, cuts both ways: we should neither dismiss a computer’s conclusions out of hand nor accept them simply because a machine produced them. Computers organize information according to selected rules and assumptions, but their conclusions still deserve scrutiny.
The wizard behind the curtain: Mr. Al Gorithm.
How does the opposite of the UNIVAC case often happen? Modern search engines bring an important question into everyday life: does appearing first make an answer more trustworthy? A ranking is not a verdict on truth, yet prominence can easily be mistaken for importance, and importance for credibility. So, the answer to the question is that most people reflexively trust the highest-ranked answer, often without double-checking or verifying it.
The issue becomes even more pressing when computers select not just the order of our answers, but the information we encounter in the first place. Today, recommendation systems reach billions of people, selecting news, videos, products, and other content. Their influence is greater because the selection is personalized and continually revised. Clicks, viewing, and purchases become signals for future recommendations, creating feedback loops that can amplify existing biases and reduce variety. What catches our attention helps determine what we encounter next.
The danger is not that every algorithm lies, but that the slice of the world it serves us can look like the whole picture, when it is, in fact, merely a reflection of our preferences and prejudices. Algorithms can help us unearth useful information at the speed of light, saving time and energy. The trouble begins when we mistake their selections for an impartial account of reality.
How do we avoid falling into this trap? The answer is straightforward. We simply follow the old Russian saying that was popularized by the Bolsheviks after the revolution that American scholar Suzanne Massie taught Ronald Reagan. Reagan, in turn, popularized by using on several occasions during nuclear disarmament discussions with Soviet officials: “Доверяйте, но проверяйте.», “Trust, but verify.”
Charles A. Ray served 20 years in the U.S. Army, including two tours in Vietnam. He retired as a senior US diplomat, having served 30 years in the U.S. Foreign Service, with assignments as ambassador to the Kingdom of Cambodia and the Republic of Zimbabwe, and was the first American consul general in Ho Chi Minh City, Vietnam. He also served in senior positions with the Department of Defense and is a member of The Steady State.
Founded in 2016, The Steady State is a nonprofit 501(c)(4) organization of more than 450 former senior national security professionals. Our membership includes former officials from the CIA, FBI, Department of State, Department of Defense, and Department of Homeland Security. Drawing on deep expertise across national security disciplines, including intelligence, diplomacy, military affairs, and law, we advocate for constitutional democracy, the rule of law, and the preservation of America’s national security institutions.
