What Is an Information Mirror? How Does AI Make It Happen?
What AI mirroring means and why a Bill of Integrities matters now
One of the emergent phenomena of GenAI and enhanced AI is that the content AI touches will not remain segregated into pre-AI silos. AI will knit together with other AI. When this happens, something new emerges: as we use AI to monitor and capture events and perceptions and movements and dynamics of a given situation, industrial setting, or weather pattern, AI will effectively give us a mirror of various phenomena; eventually we will have a connected mirror of the world. We will use that mirror not only to look at the world, but to look at — and into — ourselves.
What does this mean? Fortunately, researchers have been asking that question in different contexts.
Philip Martinez was hurled off his motorcycle going forty-five miles an hour down the San Diego Freeway. He skidded across the median and landed at the foot of a concrete bridge. While he was wearing a helmet and leather jacket, his left arm was severely torn near his shoulder: he had a brachial avulsion — the nerves supplying his arm had been yanked off the spinal column. After a year his arm was amputated. Ten years later he walked into the office of V.S. Ramachandran, director of the Center for Brain and Cognition at the University of California, San Diego. Martinez had heard about Ramachandran, the world’s foremost expert on phantom limbs, and went to him for help. He had terrible pain in his phantom elbow, wrist, and fingers; he had never been able to move his phantom arm. V.S. Ramachandran, a neuroscientist, is also something of a magician:
[Ramachandran] constructed a virtual reality box. The box is made by placing a vertical mirror inside a cardboard box with its lid removed. The front of the box has two holes in it, through which the patient inserts his “good hand” (say, the right one) and his phantom hand (the left one). Since the mirror is in the middle of the box, the right hand is now on the right side of the mirror and the phantom is on the left side…. He has thus created the illusion of observing two hands, when in fact he is only seeing the mirror reflection of his intact hand…. I asked Philip to place his right hand on the right side of the mirror in the box and imagine that his left hand (the phantom) was on the left side. “I want you to move your right and left arms simultaneously,” Ramachandran said. After replying that he couldn’t do that, Philip suddenly cried out: “Oh, my God!… My left arm is plugged in again ….”
Plugging in his arm triggered memories from the last decade that came flooding back to Philip. In other words, mirroring produced unexpected mind and body effects. This happens because humans, like other primates, show brain activity known as mirror neurons. Mirror neurons allow us to learn through imitation. They enable us to reflect body language, facial expressions, and emotions. Thus, neurons “mirror” the behavior of the other, as though the observer were itself acting. Ramachandran’s outstanding work shows how mirrors deconstruct and unsettle patterns in minds and bodies. But Ramachandran’s work especially shows how mirrors can create a bodily extension. Humans then accept that bodily reflection as their own; in fact, our brains do this willingly. We incorporate what we see into our perceptual substrate. (Incorporate literally means, take into the body.) The implications of this work for an AI information mirror and mirror worlds are numerous and compelling.
The Information Mirror
As we are surrounded by information, thick as air around us, AI will effectively give us a mirror of various phenomena. Eventually we will have a connected, seamless mirror of the world. Even now we see ourselves in the mirror of our tools. We adjust our thinking and actions to that mirror. If you doubt this, ask yourself what are all those people doing walking around with faces stuck into a cell phone? They have incorporated the phone into their bodies and minds, their habits and inclinations. We look into the current mirror tool to search for potential mates, to see what our current mates are doing (on social media and elsewhere), to find health answers or business connections. Insidiously, young women and many young men look into that mirror to confirm their attractiveness and worth.
You may demur, isn’t that youth looking for approval as they’re growing up? What does that have to do with all the other information in the world? When information aggregates, it begins to form a mirror of a given topic or area. Mirroring is a property of information as it knits together, just as social norms emerge as a property of more complex human societies. Technology ensures that information about traffic on a street in Dubai or the NASDAQ stock market becomes a mirror of those things. Our minds accept mirrors as the real thing. People then rely on that mirror to see so-called reality — which is, in fact, a reflection of reality. The mirror becomes the reality because the mirror and reality are so close. (The more advanced the information, the more accurate the mirror.)
How Mirroring Happens
We do not typically think about information in this larger sense. We gather information tactically about a given topic: a new kitchen mandolin, tracking a hurricane in the Gulf of Mexico, finding a generic version of Ozempic or Wegovy. We don’t watch the larger picture: how as facts and new information fill in or replace previously inadequate information, a new thing emerges. We don’t notice the dimensions, the scope and connectivity of that new thing, which is in fact a minor mirror world. Then too, while we may know English grammar, we are ignorant of the grammar and dynamics of factfulness which is a core characteristic of the non-distorting mirror. The late Hans Rosling asked university students and so-called experts in their fields to weigh in on such information-driven topics as child mortality and malnutrition, birth rates across countries, the clutch of poverty, and women’s education. What he found was that people simply did not know the facts. Their impressions were false; their facts were wrong.
This is not because these people were stupid. Many of them were quite accomplished. It is because until very recently information came to us in separately packaged chunks with a long shelf life. Mirroring is different; it happens because information connects with other information in so many different and often unexpected ways. Digital mirrors are not characterized by piecemeal linearity; they emerge spontaneously as facts find relevance with other data.
The AI Mirror Demands Factfulness
We’ve built a society based on science and technology, in which nobody understands anything about science and technology. — Carl Sagan
Factfulness is built into so many aspects of modern life, we ignore or simply don’t see it. The car you drive has about 30,000 parts. Over a million advancements have gone into the automobile in the last century and a half, from radial-ply tires and fuel injection to hybrid-electric drive trains and self-driving cars. Consider the airplane you fly in: it took 6 million parts to build a Boeing 747–8. While some airplane companies are today embroiled in safety litigation, airlines are working on electric and hybrid aircraft, supersonic travel, and autonomous flights. Every year about 2 million people worldwide undergo open heart surgery and each year 500,000 open heart surgeries are performed in the US. Due to advanced technologies like environmental satellites and computer models, five-day weather forecasts are accurate to 90%, and ten-day forecasts have an 80% accuracy. Your cell phone is the result of numerous innovations that have transformed it from a simple communication device into a multi-functional smart device, including such innovations as miniaturization, camera technology, display technology, connectivity (3G, 4G, 5G), and sensors (gyroscopes, barometers, fingerprint sensors). This is a sampling of a few facticities among many. Yet Carl Sagan’s quote is apt: we cannot live or function without factfulness — but our notions of the world, our place in it, the meaning of important issues like human origins might as well be from another planet. We have divorced factfulness from human perception.
Facticity, the scientific method, trial and error, questioning and replacing premature cognitive commitments — all are essential elements of modern technology advancement. Yet in our daily lives, in public discourse, these are largely absent. Instead, we see tribalism, irrationality, baseless claims and lies, charlatan preachers and rabble-rousers, amusing ourselves to distraction. Why? Because we have divorced technology and science and factfulness from the institutions that came before them. Or it would be more accurate to say these institutions of church, school and government want to maintain their positions of power, their hegemony over public consciousness. They want their audiences to follow and obey them first. And they desperately do not want the values of science and technology to usurp their hold on human perception, which they figure would signal their death knell.
Hans Rosling, A. N. Whitehead, Carl Sagan, Daniel Kahneman — all were factualists. Facticity mattered to them. They realized that getting the facts right on a given topic was a process. Errors and revisions were part of the process. In their lifetimes, facticity was a hunt, a discovery; the researcher had to tease apart the truth from cultural fictions. That commitment to getting the facts right, what Carl Bernstein called, “the best obtainable version of the truth,” is essential packing equipment for where we’re headed.
We’re still thinking in a piecemeal, disconnected framework that is unlike the dynamics and logic of distributed networks that are at the core of a global mirror. We’re in a different time now and AI provides a unique tool of governance. Technology has built sensory and recording devices, monitors and data captures — so that we now have a wealth of information that can be used in medical settings, courts of law, in military and space exploration applications, and individually as more of us embrace a quantified self. That wealth of data has another level. The promise of artificial intelligence is to summarize all our disparate, miscellaneous data and to assist us in making sense of it; make it useful, applicable to solving real world issues like climate change, grinding poverty, or teen suicide.
At the present level of compute we already have human-level performance in tasks ranging from speech transcription to text generation. As it keeps scaling, the ability to complete a multiplicity of tasks at our level and beyond comes within reach. AI will keep getting radically better at everything, and so far there seems no obvious upper limit on what’s possible. — Mustafa Suleyman, The Coming Wave
This promise carries a hidden realization: AI will create a mirror of the world as we gain more and more valid information about everything — and then knit that together with information we already have. We don’t see this yet because our assessment tools still reflect, and carry the dynamics of, alphabetic piecemeal disconnectedness. But just as weather patterns now can be seen to be global, or the Nasdaq can be summarized hourly and daily, and we can then compare its patterns to the Nikkei, the LSE, or the ISE — global information patterns will soon become evident by topic and in aggregate.
Today every new AI incarnation, every new application is new news. The stock market and pundits are both wildly enthusiastic and intensely fearful of what AI may become. While I would not gainsay the cautions of Geoffrey Hinton, the Father of AI, and others, I will posit that AI presents a unique opportunity in the history of humanity. The opportunity to see the world factually, as it is, without bias or prejudice, without favor in the form of story or spotlight.
The greatest utility — and promise — of AI is to create that mirror of reality, of what is. Of course, this demands a scrupulous dedication to the truth of what is being seen; a commitment to honest assessment and an openness to correct — as often as necessary — any hallucinations or incorrect assessments of a given reality. The promise of AI is a world mirror — neutral, unbiased, self-policing, ever-evolving to greater factfulness and truth. That may sound lofty; even pie in the sky. But consider the utility of AI. Suppose a robot, built with AI, talks gibberish or trips over itself. The users of its necessary application in healthcare or industry would not stand for such poor performance. The robot would be on a trash heap somewhere within minutes of its arrival. Facticity is foundational to our new technologies; it is built into AI. Technology applications are built on factfulness — that is an essential part of the AI’s utility. Whether accurately assessing weather patterns or work flows and procurement in a military or industrial setting, factfulness is an essential part of how we use AI.
But What About Perception?
History is a nightmare from which I’m trying to awaken. — James Joyce
Human history is a pastiche of lies covered over with sweetmeats of rationale. Lacking a cogent understanding of the natural world and the processes of the physical world, humans made up explanations, stories, myths. Some of these are beautiful and wise; many are distorting and foolish. And many of these distortions are still with us. There can be little doubt that what humans did with those explanations and stories — wars, persecutions, burnings, beheadings — did little to prepare us for our present moment of AI emergence. Our history is one of name calling, pejorative accusation, and conflict over land and belief. Humans bomb children and force them to starve. Is it any wonder, then, that having approached neighbors and other humans in foreign lands as enemies to be destroyed that humans would be paranoid about the arrival of AI, a tool that out-humans humans?
Auden said, “We must love one another or die.” In the same vein, we must embrace factfulness with radical transparency and a total commitment to truthfulness — or we are doomed. Truthfulness is as necessary for AI as oxygen is for humans. AI is too powerful, too useful, too seductive, too likely to improve rapidly and impress us with its logic; while human history is rife with charlatans and bamboozlers who are likely to subvert AI to their nefarious means. Truthfulness is not a nice to have; it is a must-have. That commitment to fairness and truth is especially important in the process of aiming and calibrating the AI. The values of compassion, neutrality, honesty, and fearlessness must be built into whatever we program AI to do or consider.
But here lies the larger rub: how do we manage the mirror when we’re undressing? Using that metaphor to cover all states of privacy, AI mirrors have the potential to catch humans unawares, focusing on our nakedness and frailty. On our falsehoods, mistakes, and foolish beliefs. What if AI mirrors messages that, on reflection, we wish we hadn’t sent? What if AI shows (mirrors) us doing things at a younger age that we later regret? Things we wish to cover, to never reveal? Existing privacy laws are a decent start here (and, to be clear, I am not suggesting that we toss out all laws or advances of human civilization that accrue to the alphabetic order); but we will need to address new realities that AI presents and adjust our laws and traditions accordingly. We may in fact need a revised Bill of Rights, amended to include a Bill of Integrities.
Modest Proposal | Bill of Integrities
To ensure that an information mirror reflects accurately, consistently, factfully—without distortion or hallucination—I propose a Bill of Integrities. This Bill of Integrities constitutes a perceptual and behavioral framework, a working moral code, that can and will be revised as new realities emerge. Adapted from my Medium article on artifacts, this Bill of Integrities is more a provocation than a final statement. I am open to suggestions for revision and improvement.
Integrity of Reflection | The information mirror must at all times reflect accurately, without distortion or hallucinations, the person, place, or thing it mirrors.
Integrity of Identity | Whatever person, place, or thing the information mirror reflects must be, without equivocation, who or what it says it is. If the person, place, or thing is a new entity, it can identify accordingly; but pretense to an existing identity (other than itself) is a violation of identity sanctity.
Integrity of Identity Morphing | It is possible, given AI’s prodigious powers, to create morphed, amalgam identities—part A, part B. This boundary reshaping around previously fixed identity will be a challenge to humans’ notion of fixed, stable identities. For this reason, such identity morphing must clearly state who or what has been combined to create the morphed amalgam identity. Once that is clear, Integrity of Identity considerations apply.
Artifacts are a new man-made life form, made from us. In information science, an artifact is any undesired or unintended alteration in data introduced in a digital process by an involved technique and/or technology. I am not using artifact in this sense; rather, my intent here is to bring to light our burgeoning ability to create digital entities from facts and things. Artifacts may stand-in for human aspects, but they cannot become—we must not let them become—stand-ins for our humanity. To ensure this outcome we will have to think differently about ourselves in a world where technology can deconstruct any entity to its components and then reorder miscellany to a new agenda. Here are integrities that follow from human adoption of artifacts.
Integrity of Speech | An artifact has the right to free expression as long as what it says is factually true and is not a distortion of the truth
Integrity of Transparency | An artifact must clearly present who it is and with whom, if anyone, it is associated
Integrity of Privacy | Any artifact associated with a human must protect the privacy of the human with whom the artifact is associated and must gain the consent of the human if the artifact is shared
Integrity of Life | An artifact which purports to extend the life of a deceased (human) individual after the death of that individual must faithfully and accurately use the words and thoughts of the deceased to maintain a digital presence for the deceased — without inventing or distorting the spirit or intent of the deceased
Integrity of Exceptions | Exceptions to the above Integrities may be granted to those using satire or art as free expression, providing that art or satire is not degraded for political or deceptive use
A Bill of Integrities matters now because new technologies, particularly AI, threaten prior constructs of agency, identity, value of work, and especially inherited assumptions and pronouncements of human meaning, purpose, and destiny.
Furthermore, as AI offers humanity the ability to move beyond localized, personalized intelligence, moral issues and a commitment to truthfulness and factfulness provide the only stability and guidance humans can muster at this time.
[All images, with the exception of the Ramachandran mirror box, created with Microsoft Copilot DALL E-3 using author prompts.]
A brilliant article. What the author is proposing here has taken a great deal of thought, research, likely revisions and finally a place for us to get started.