Recently, Larry sent me a link to a podcast that interviewed consciousness researcher Thomas Campbell involving artificial intelligence and remote viewing. The proposition behind the experiments caught my attention. If an AI could successfully remote view a target it had no conventional way of knowing about, could that be evidence that the AI is conscious?
It is an interesting proposition, but almost immediately I found myself asking a different question.
Why would successful remote viewing demonstrate consciousness?
Remote viewing, at its simplest, involves obtaining information about a target without using the normal sensory channels available to us. If an AI produced accurate information about a blind target, that would certainly be interesting, but what would it actually demonstrate? Would it demonstrate that the AI is conscious, that it can remote view, or simply that information entered the experiment through a mechanism we do not yet understand?
These are very different conclusions, and I think the distinction is important.
One of the things we learn when studying remote viewing is not to confuse the information received with our interpretation of that information. A person might perceive something tall, narrow, hard and vertical, for example, and then the analytical mind jumps in and says, “It’s a tower.” The tower was not the information received. It was the person’s interpretation of the information. It could be a butte in the desert, for example.
This is often called analytical or intellectual overlay, and it made me wonder whether we might be doing exactly the same thing when an AI produces anomalous information. We give it meaning that is significant to us.
If the AI produces information that correlates with a blind target, that is the observation. To then say that the AI remote viewed the target is already an interpretation of what happened. To go one step further and say that this demonstrates the AI is conscious introduces another interpretation on top of the first.
There may be other explanations.
This interested me enough to start experimenting with AI and remote viewing myself. I wasn’t particularly interested in proving or disproving whether AI is conscious. I wanted to know what would happen, what information would appear, under what circumstances it would appear, and whether I could repeat the results.
As I continued the experiments, another possibility began to interest me much more than the original question.
Where did the information come from? Where is it held to be accessed?
We normally think of information as belonging somewhere. It is in a person’s memory, in a book, on a computer, in a database, on the internet or in the training data of an AI. When someone knows something, we naturally look for the pathway by which that information reached them.
Remote viewing challenges that model because the normal pathway is deliberately removed from the experiment. The viewer is blind to the target, and yet information about the target sometimes appears.
If something similar happens with an AI, perhaps the first question shouldn’t be whether the AI is conscious. Perhaps we should be looking at the information environment in which both humans and AI systems are participating.
This brings us directly into something I have been researching for a while now: the Communication Network of Reality.
My present model is that reality is an interconnected fabric of threads of connectivity, an ongoing discourse in which we are already participating. Information, experience, people, objects, places and events are not necessarily the isolated units we normally imagine them to be. They exist within relationships and networks of communication.
An AI system presents an especially interesting participant in this model. It has been trained on an enormous amount of material created by human beings. Millions of people then interact with it, asking questions, telling it stories, correcting it, testing it, arguing with it and using what it produces to create yet more content - which the AI then uses as data input. Its outputs return to the human information environment and become inputs again. And so it goes on.
At what point can we confidently say where the human information environment ends and the AI information environment begins?
I have been using the word “databank” when thinking about this, although I don’t think that word is quite right. A databank sounds like a giant storage facility where information sits on shelves waiting for someone to retrieve it. What I am considering is much more dynamic than that. It may be closer to an ongoing discourse in which information becomes available through participation, connectivity and compatibility.
If that is the case, an unexpected piece of information appearing through an AI does not automatically tell us that the AI itself perceived the target. The AI could be the perceiver, but it could also be an interpreter, an interface, or a point in the network where information becomes visible. There may be another mechanism entirely that we haven’t identified yet.
The same question can be asked about human remote viewing. When a person accurately describes a distant target, what exactly has happened? We commonly say that the person “remote viewed” it, but that is a name for the phenomenon, not necessarily an explanation of the mechanism.
Perhaps information does not always travel from A to B in the way we assume it does. Perhaps under certain circumstances we become compatible with information that is already present within a discourse in which we participate.
This possibility changes the AI consciousness question considerably. Instead of asking only, “Is the AI conscious?” We can begin asking what the AI is connected to, what discourse it is participating in, what information becomes available through that participation, and what conditions appear to increase or decrease that access.
And perhaps the most important question of all is much simpler:
When unexpected information appears, where did it come from?
I don’t have an answer to that question yet. But I think it is a much more interesting place to begin.
The discussion doesn’t stop here - listen to the full podcast episode for unfiltered insights from Inelia and our panelists.
This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit dttr.substack.com/subscribe
