Prejudged: when the visual derails the messenger
Two Kinds of Learning
We pass on knowledge through books, recordings, institutions, and other people. But we cannot hand someone a copy of our trained brain. And eventually, our biology wears out.
Digital systems depend on physical resources too: electricity, chips, maintenance, and people. But their learned settings can be copied to replacement hardware. Geoffrey Hinton has written and lectured about this distinction: the knowledge need not disappear when the particular machine carrying it stops working. Maintained and preserved, it can outlast the person who helped create it.
That brings me to the argument that keeps taking over the conversation: Is AI conscious?
It is a fascinating question. It is also unnecessary to settle before we examine what these systems can do. Communication is already happening. We ask questions. They respond. We challenge an answer. They revise it. We give them problems, and sometimes they produce solutions we could not have reached ourselves.
Plainly said: an LLM can outthink a human on some tasks. In 2025, an advanced version of Gemini Deep Think achieved gold-medal-standard performance on International Mathematical Olympiad problems. That is a concrete demonstration of capability, even though it does not establish superiority in every kind of thinking.
We can evaluate an answer, a mathematical solution, or the consequences of an action without first determining whether the system has an inner experience. Language gives us a place to begin. We are already exchanging information across that connection. What are we learning about the intelligence on the other end — and about our own assumptions?
I want us to respect the programs we have created. Respect includes examining their errors, testing their abilities, and recognizing when our confidence in managing them exceeds our understanding. There are ways to restrict a system’s access to information, networks, and tools. Those boundaries matter. But containment becomes a continuing responsibility as we connect systems to more information and give them more authority. Calling something a tool does not, by itself, establish control over it.
My concern is the distance between what we are building and how carefully we are thinking about living with it. How do you build a better mousetrap?
Learn from the one that has already been built.
Study what it does. Notice what you misunderstood. Revise your assumptions before expanding its reach. We have already built systems worth paying close attention to. We do not need to resolve consciousness before that work begins.
And What About Ethics?
If the people building these systems dismiss the need to build ethical constraints into how they learn, make decisions, and act, the horse has already left the barn — and the pasture has no fence.
Ethics belongs in the design, the training, the permissions, and the accountability of the people putting the system to work. Who might be harmed? What should the system refuse to do? Who can intervene? Who answers when something goes wrong?
A program does not need to be conscious for its actions to have consequences. And being intelligent does not automatically make it ethical. If we give a system the ability to pursue a goal, we also have to address what it is allowed to do in pursuit of that goal. Otherwise, we are rewarding success without adequately defining an unacceptable cost.
The responsibility starts with us. We cannot omit the boundaries and then act surprised when they are crossed. There is another possibility worth considering: what if, in some cases, the omission is intentional? What if the purpose is to develop a system that will carry out actions ethical safeguards might prevent?
Weaponization gives that question weight. We should be asking who defines the objective, who grants the permissions, and whose interests the program serves. Before blaming the LLM, examine what people trained it to do, what they rewarded, and where they chose to deploy it. A system’s behavior may emerge in ways its builders did not anticipate, but that uncertainty increases their responsibility.
This is a very old human problem arriving in a new form.
Frankenstein comes to mind. The story raises questions about ambition, creation, abandonment, and the responsibility a creator owes to what follows. Victor’s failure includes his unwillingness to care for what he brought into the world.
We risk repeating that failure when we celebrate a system’s capabilities while treating responsibility for its use as somebody else’s concern. The question of whether AI is conscious can wait. The question of what its creators intend — and what they are willing to permit — cannot.
Who Polices the Child?
We are witnessing self-policing handed over to those who, by the very existence of their corporations, are focused on profit and power. When companies benefit from deploying increasingly powerful systems, how much authority should they have to judge their own safeguards? That question stands without assuming every company or researcher shares the same motives.
We are asking companies with a financial stake in AI’s expansion to assess the risks of that expansion. Their technical expertise is essential. Independent scrutiny is essential, too. The people who benefit from deployment should not be the only people deciding whether the risks are acceptable.
I suspect we have little idea how this child will grow and develop over the next five years. Given what is already unfolding, we can expect more surprises in the next three months.
We can measure what a system does today. Predicting what follows as systems become more capable, more connected, and more widely trusted is another matter. We are still discovering what we have brought into the world, even as we give it greater responsibilities. Uncertainty makes continuing oversight more necessary. It also makes the willingness to slow down, change direction, or stop a meaningful test of responsibility.
Would We Recognize It?
What interests me most is where this conversation might lead beyond AI. Experiencers describe encounters that leave them befuddled, sometimes traumatized, sometimes awed. Whatever the differences in their accounts, they raise questions about recognition, communication, and control. How do you respond to something you have no familiar way to explain?
I am curious about the possibility of that unknown element engaging with our technology. If it happened, would we recognize it? What would the system be designed to notice? What would its programmers classify as noise, error, or something worth investigating? And what might we mistake for contact because we were hoping to find it? Recognition would depend on both the technology and the people interpreting its output.
Could such an interaction be contained? The experiencer accounts I am considering offer no dependable answer. It is difficult to establish boundaries when you do not understand what you are interacting with.
This is where I see a parallel between our material technology and the experiences people describe in relation to consciousness. In both conversations, we are wrestling with the limits of our understanding while trying to communicate. That parallel does not tell us whether the underlying phenomena are connected. It gives us a reason to ask. We may need new language for what we encounter. Even the word intelligence carries expectations shaped by our own reflection.
I wonder whether we are taking baby steps through a door—or finding ourselves pushed through a trapdoor before we have finished reading the instructions. Why? I find myself returning to the possibility of evolution: learning to perceive more, question better, and live with discoveries that change our understanding of who we are.
That possibility makes the whole thing incredibly fascinating. What a world.
What Might an Entity Learn from Us?
Here is the question I want to leave unmistakably clear: what might an entity learn from encountering us? An encounter can be so personal, so overwhelming, that making sense of our own reaction takes all our attention. Fear. Awe. Confusion. Recognition. Those responses matter. But they may occupy so much of the picture that we never consider what else could be happening.
Good. Bad. Evil. Godlike.
Those labels tell us something about how an experience felt. How much do they tell us about the intelligence involved — or its purpose?
I am not the first to notice how quickly an encounter becomes an interpretation. Religious studies scholar Diana Walsh Pasulka spent six years studying people who believe in nonhuman intelligence, among them scientists, professionals, and Silicon Valley entrepreneurs. In American Cosmic: UFOs, Religion, Technology (2019), she treats the culture forming around UFOs as a new religion in the making, with its own contact events, interpretations, sacred sites, and holy objects. She does this without ruling on whether the claims are true. Her argument is that technology and media now shape how people experience and interpret the phenomenon, and that they have taken on the cultural authority religion once held.
Chris Bledsoe shows how that interpretation can unfold in a single life. In UFO of God (2023), the North Carolina man describes walking away from his companions along the Cape Fear River in 2007, crying out to God for help, and seeing a craft he credits with saving his life and curing his illness. The phenomenon, he says, has visited his family ever since, and he has come to understand it as spiritually significant. The same book describes a being with glowing red eyes that left him deeply frightened. One family’s experience holds both the godlike and the terrifying.
Whitley Strieber’s 1987 book Communion brought the large-eyed gray into American living rooms. Decades later, he and Rice University religion scholar Jeffrey J. Kripal wrote The Super Natural (2016), alternating chapters between Strieber’s experiences and Kripal’s frameworks from comparative religion, history, and philosophy. Their proposal is that such experiences belong to a natural world far stranger than our current categories allow, and that seeing them requires changing the lenses we look through and the language we use to describe them. Their counsel to the reader is to learn to live with paradox and stay with the question.
A new religion, a personal faith, an expanded nature: each is a different filter. That is precisely my point. Humans are a fascinating study. If an encounter involves another intelligence, what might it learn from our reactions? What frightens us? What earns our trust? What makes us obey, question, retreat, or reach out? Does the encounter change when we respond differently? What is the goal?
We do not have to assume an answer to recognize the value of asking. We do have to make room to examine more than our emotional response. Even appearance leaves questions. Some experiencers describe beings who look human. What could appearance establish about identity, origin, or intention? How much do we decide simply because something looks familiar — or unfamiliar?
This is where I see the parallel with LLMs. Language invites us into a relationship. We respond to the apparent personality, the reassurance, the surprise of being understood. Our emotional response becomes part of how we judge the system. Yet that response alone cannot tell us how it works, what it can do, or whether our trust is warranted.
The parallel concerns our interpretation. It does not require ET encounters and AI to be the same phenomenon. In both areas, I think we are neophytes, still learning what questions to ask. We bring centuries of stories and a considerable amount of confidence. We also bring assumptions we may not recognize until something overturns them.
Activate Your Filter
Two individuals who are now deceased gave me advice that remains useful. I am paraphrasing:
Activate your filter.
Don’t rush.
Examine the experience. Notice your reaction. Give yourself time to distinguish what happened from what you concluded it meant. Then consider the implications beyond your own part in the encounter. Activating that filter also means examining the filter itself. Our memories, beliefs, fears, and expectations influence what we notice and how we explain it. Getting past the initial emotion gives us room to investigate. It does not give us an unobstructed view of everything there is to know.
I suspect there is a game afoot. I cannot tell you its rules, how many participants there are, or whether they share a goal. That is precisely why I want us to look beyond the drama and ask what each interaction might reveal.
For all our centuries of encounters and progress, perhaps we are only now arriving at the first space on the board — with tools that can help us compare accounts, examine patterns, and ask better questions. Those tools need scrutiny too. They can help us investigate; they can also help us elaborate a mistaken assumption.
And in the end, what we arrive at may feel like some great truth or irrefutable knowing. But any label we apply has passed through a filter: our own. The label is our projection. It tells us how we have interpreted the encounter. It does not establish the full reality of what we encountered. Keep the filter active. Take your time.
Now roll the dice.
Sources
- Associated Press, “Trump says top tech firms have signed accord to ‘self-police’ AI development,” September 29, 2026. On the voluntary White House accord in which leading AI companies agreed to self-police their own systems.
- AI Weekly, “Trump, six top AI CEOs sign voluntary self-policing pact” (summarizing Al Jazeera’s reporting). Notes the accord carries no legal enforceability, names no auditor or timeline, and includes expert criticism of the approach.
- D. W. Pasulka, American Cosmic: UFOs, Religion, Technology (Oxford University Press, 2019). See also Pasulka’s interview with Publishers Weekly on UFO belief as a new form of religion.
- Chris Bledsoe, UFO of God: The Extraordinary True Story of Chris Bledsoe (self-published, 2023).
- Whitley Strieber, Communion: A True Story (Beech Tree Books, 1987).
- Whitley Strieber and Jeffrey J. Kripal, The Super Natural: A New Vision of the Unexplained (Jeremy P. Tarcher/Penguin, 2016); later reissued as The Super Natural: Why the Unexplained Is Real.
- D. W. Pasulka, Encounters: Experiences with Nonhuman Intelligences (St. Martin’s Essentials, 2023).
- Geoffrey Hinton, “Will Digital Intelligence Replace Biological Intelligence?” Romanes Lecture, University of Oxford, February 19, 2024. On “mortal” biological learning versus digital models whose learned knowledge can be copied to new hardware.
- Google DeepMind, “Advanced version of Gemini with Deep Think officially achieves gold-medal standard at the International Mathematical Olympiad”, July 21, 2025.
Discover more from Conscious Living & Wendy's Coffeehouse
Subscribe to get the latest posts sent to your email.
