Human Being, Human Doing, Robot Being, Robot Doing. Homo Sapiens, Robo Sapiens. What might artificial life forms do for us, biological humans, to satisfy the psychological need we have for being together?
In Shoji Morimoto’s memoir, ‘rental person who does nothing’, Shoji aims to be with another whilst doing little. Shoji offers himself as a rental person to be spatially and socially with others in various situations, engaging as a presence, but not doing any tasks. He gives examples of listening without offering any advice but he acknowledges that he is listen-ing. He will sit & wait under a designated blossom tree to save a spot for a blossom party but will not pick the spot for the client. He agrees to be in a park with a person who says hello to strangers while wearing an animal costume. The clients value Shoji’s presence as a catalyst for their own actions and to aid their life journey. These meetings are one-off and not repeated with the client.
“I’d like the world to be one where where even if people can’t do anything for others, they can still live stress free lives. This is very important to me because of the gap that exists between the value that I sense in people and the value assigned to them by society.” Shoji Morimoto
Why would a person engage a rental person who does nothing? Yuval Noah Harari’s Sapiens thesis suggests an answer: we are the only animal that can cooperate flexibly in large numbers. Like a pet, another human being who is willing to be with another in a non-judgmental and non-threatening way, holds a strong attraction as a witnesses of another being’s existence. More socially-isolated selves find that problems and tensions can only be solved by being witnessed by another being. A witness can “see or know by personal presence”.
To extend the Sapiens thesis, are we are the only animal that can cooperate flexibly in large numbers with both animals and artificial beings? that is, with matter that we have animated into various forms of consciousness, with variable levels of intelligence and variable human relate-ability. We already gender our voice assistants because we can relate to tonalities like gender, age and accent, by displaying some of the characteristics of a relatable being. We don’t choose Dalek voices for our companions.
We are close to a future where Artifical Friends (AFs) are readily available as voices, robots and copies of deceased human beings. AF is a term from Kazuo Ishiguro referring to Klara, the hero of ‘Klara and the Sun’. Can my being (as a human) be witnessed by an AF? Can an AF witness another AF’s (robo) being? Can a group of AFs provide community for a human? In a Quaker meetinghouse, “where two or three are gathered together in my name, there am I in the midst of them”, is there communion when an AF is present as a participant? While a human is dying, does a robotic hand holding the human’s hand, and the human’s confession to the AF of a deep-seated regret, create the same solace in the human as a human hand and human witness?
We expect robots will do things for us, like vacuuming and washing dishes, and will act to further human goals; it’s easy to envisage agricultural robots growing & harvesting food for humans. But what are our expectations for being witnessed by these artificial selves when we recognize them as fellow conscious beings? The book Conscious by Annaka Harris lays out the mystery of consciousness succinctly and with clarity. Its single most important insight is that the only thing we truly ‘know’ is that we are conscious. And the second astounding insight, that the scientific method says nothing about the intrinsic nature of matter and energy; science only predicts what energy and matter will do.
The Hard Problem of Consciousness is paradoxical and is perhaps unsolvable— how is it that stardust, mysteriously organized into distinct elements and molecules, can accumulate into forms that have an experience? — it feels like something to be an amoeba, a bat, or a human. Is it like something to be a dishwasher or to be a web server or to answer humans requests as a Large Language Model?
What did it feel like to be the re-animated Frankenstein made up of multiple human parts? Frankenstein was a biological reanimation — but digital animation of a human-like consciousness is something novel. Mary Shelley’s Frankenstein is a fictional character and as such is also an imagined consciousness. Some authors’ characters like Sherlock Holmes or Kay Scarpetta are so well realized as to be repeatable and predictable in multiple contexts — one could predict how they might react or speak in a new situation.
Panpsychism seems both defensible and reasonable, per Harris’ Conscious book; a theory that all matter is imbued with consciousness. Almost immediately one has to define terms carefully and get away from a common-sense definition of consciousness — the keyboard I am typing on is unlikely to be conscious in any way that is relatable to us humans; consciousness is more like one of the basic forces, if using a panpsychist definition. We don't really know what matter and energy are, but we know that we experience energy and matter thru our bodies and via thought.
If we attempt to create a form of pseudo-human consciousness in a physical computer, and panpsychism is a real phenomenon, arguably we are mixing two forms of consciousness in that device; and perhaps may be surprised by the admixed consciousness results.
Reading the paper, Is Artifical consciousness achievable? Lessons from the human brain by Farisco, Evers and Changeux, April 2024, the authors reinforce the view that the term consciousness is misleading and that a new term might be used for denoting artificial or machines-based forms. They also propose that we develop precise language to describe the various dimensions that consciousness takes and to measure the level of consciousness that a system (or person) may be capable of displaying.
Concerning the conceivability and feasibility of developing artificial consciousness, we will distinguish between:
(a) the replicability of human consciousness (which we exclude, at least in the present state of AI development, a stance which is scarcely controversial)
(b) the possibility of developing an artificial conscious processing that may bear some resemblances but still is profoundly different than human (which we do not exclude in principle, but consider difficult to prove for both theoretical and empirical reasons).
The authors argue that human brains develop due to spontaneous activities which modulate our sensory perceptions and result in cultural evolution associated with three key areas: “rational thinking (science), rules of conduct (ethics), and shared feelings (art).”, and that “AI lacks an evolutionary development strictly homologous to that of the human brain even networks that learn to control a robotic agent have been designed to follow a path of ‘synaptic’ growth.”
Using evolutionary language based on development in newborn humans, they define stepwise conscious processing as:
- Minimal consciousness — can display spontaneous motor activity, and can store memories long-term. The mouse remembers seeing a cat. A 22–30 week human fetus is at this stage.
- Recursive consciousness — can make functional use of objects and can point at things, an;d are capable of elaborate social interaction and shared attention (OMG, I think there’s a leopard near our Chimpanzee family. There!). Newborn humans are here too.
- Explicit self-consciousness — two year old infants self recognize in mirrors and have basic language. Chimps can also reach this level.
- Reflective self-consciousness — a full conscious experience with ‘first person ontology and report ability’ available to humans after 3–5 years of age.
These are helpful terms to describe relative levels of consciousness in an artificial system, but that level descriptor can imply that the artificial consciousness will not ascend levels, as we do, given the varied form of evolution that it will follow, not ours. Our evolution has enormous social relevance, not accessible to the artificial consciousness in the same way especially in the emotional realm tied to a gendered biological body and brain, and to the ethical and artistic realms.
While pondering this topic, it occurred to me that there is no specific reason why a household AF robot should present the same personality to all its human clients. We too, will present different aspects of our personality to others and behave quite distinctly in professional and friend contexts or in group settings. An AF could be a composite of tens of pseudo-consciousnesses if it had the computational and memory capability to do so.
We need to prepare for this future of dynamic artificial friends — the artificial friends and characters from beloved fiction books will seem less dimensional to us then. Our present language for describing human-generated alien pals and agents, is inadequate for the task. We need to describe ourselves well, in order to differentiate, assimilate and co-exist with human-like and ahuman consciousnesses. We have done this well in religious, commercial and mythological thinking for disembodied entities: we work for companies, we call for divine help, we ascribe disasters to mythic creatures.
How did the collection of stardust that is your body awaken to itself? This is a deep mystery. Can awakened matter, awaken more matter? It seems likely right now. Promethean clay taken to a new level — a truly fascinating power, not to be undertaken lightly nor without the right words to describe it.
