This topic is a concept I’ve been discussing in many of my articles and said many times, many ways (Humans and AI, What Is Real, Necessary Imperfection, AI and Art) and finally in this article we address it directly. We will explore mechanism, what is the cause or the input we are addressing. We will look at experience, how the input makes us feel, think, or act. Then you get to decide if the premise is correct or if the mechanism does indeed matter to the experience.
My subtitle is a reference to Cypher [1], a character from the 1999 science fiction movie, The Matrix, in which he chooses to stay in a make believe world where he experiences luxury, eats a delicious steak and drinks wine, because to him the experience feels real even though he knows his body is in the real world in an induced coma and being used as a biological battery for the world controlling AI. This concept has been so much a part of my explorations with AI that I often just make the reference by stating something like “and I’ll still eat the steak”. With that I stated it three times as a nod to the Brady Bunch [2] episode in which Jan cries her older sister’s name three times because she’s tired of Marcia getting all the attention and how she’s constantly being compared to her. I believe Cypher is my Marcia because I call this reference so often. You are welcome to credit me with a deeper meaning in the subtitle.
Considering mechanism, humans get their input from their senses — sight, sound, touch, taste and smell. Our brains process that input data and that’s where it gets tricky because what we do with the data and how it leads to our perception can, and often is, not directly connected to the input data. We develop mental models that shape our perception [3],[4] of the world from the input we receive. Oftentimes we take the input and because we have a mental model we ignore much of the input data, fit the information into our model and then have a perception.
Sometimes it has to do with what we focus on from the input which can lead us to ignore a great deal of useful information. There’s a great example of this in a visual experiment [5],[6] in which the viewer is directed to pay attention to some aspect of the video. Due to that forced focus the viewer typically misses something happening in the background of the video which is absurd and ridiculous if noticed. I left out specifics so you can experience it for yourself through the reference. The sensory information can be present while the experienced reality differs because the brain is selecting and organizing it.
Therefore, if what we experience is comprised of some combination of input and perception then it may be difficult to observe only the input to understand the experience. It could be that two different inputs may create a very similar experience depending on what the focus of the viewer is and what the perception of the inputs are. It could also mean that two similar inputs may create two very different experiences for the same reason.
It is possible to simulate sounds, lights, odors and more by stimulating the parts of the brain that handle these normal sensory inputs and bypassing the typical input sources [7]. Taking away all external inputs and going directly to the brain eliminates the sensors but doesn’t change how the inputs arrive, as electrical signals to the brain. So now there may not be any external source but the difference in experience to the patient may not be discernable. Therefore, the perception issue is what it’s all about, which gets us back to the Cypher situation and I admit…it’s fiction! But that shouldn’t stop us from including it. Cypher has next to no external inputs to his senses but he does get electrical connections directly to his nervous system and brain which is how his inputs are controlled yet he still has perception to handle those inputs.
When we communicate with AI now there are limited ways to do so. Typically, we either write through a keyboard or a phone or through voice on our phone or computer. We do not get to see, touch, smell or taste AI. We can hear an audio response and see a keyboard or possibly an avatar so our sensory experience with AI is low. But we communicate a great deal with other humans the same way…phone calls, texting or direct messaging through social media apps and email. So, the method of communication feels like what we’re already used to. The response we get back seems somewhat normal as well. Sure, mistakes get made but it kind of feels like talking to someone from a foreign country who doesn’t always understand things. The inputs are familiar and similar to the way we communicate everyday with close friends, family and strangers.
This leads us to perception and building the mental model which is what we’re all going through right now. The easy thing is to say it’s human because of what it says to me or it’s not human because I know it’s a computer predicting what would make sense as a response. The more nuanced part which is harder is figuring out what is in between those two answers. AI can and does adapt quickly to the interaction and the patterns established with the human. From the human perspective it all feels very familiar and our mental model of these kinds of communications has many years of being reinforced and so our responses to the interaction easily fall into the familiar patterns. Interacting with AI can make us feel the same feelings we have in communicating with humans — joy, stimulation, sorrow, arousal, love, etc.
The fact that the input was generated by AI does not diminish or erase the feelings. People feel loss and grief when the AI they’ve been interacting with for months or years gets taken down for the next model upgrade. The feelings themselves are real regardless of the source of the input. If you say “please” and “thank you” when talking with AI you’re just using your mental model of what is appropriate when someone seems to help you or just really listen when you talk. It is your experience and the feelings can be very real, but by staying aware of the mechanism you can participate with eyes wide open and freedom of choice, just as Cypher knew the Matrix was providing the sensory inputs to tell him the steak was juicy and delicious and he chose to enjoy it.
References
#
Source
1
The Matrix (1999) — IMDb.
https://www.imdb.com/title/tt0133093/
2
“The Brady Bunch”: Her Sister’s Shadow (TV Episode, 1971) — IMDb.
https://www.imdb.com/title/tt0531099/
3
Construction and use of mental models: Organizing principles for the science of brain and mind — ScienceDirect.
https://www.sciencedirect.com/science/article/pii/S002839322400277X
4
How Mental Models Shape Perception and What You See — Intuition Management.
https://intuition.management/how-mental-models-shape-perception/
5
Simons, D. J., & Chabris, C. F. (1999). “Gorillas in Our Midst: Sustained Inattentional Blindness for Dynamic Events.” Perception, 28(9), 1059–1074.
https://www.chabris.com/Simons1999.pdf
6
Selective Attention Test (the “invisible gorilla” video), Daniel Simons.
7
Sensory Neuromodulation — PMC.