What are the minimum viable requirements for a system to be self-conscious? To know that it is, and for it to have self-reflexivity.
I attended a talk by Erin Green of HeadHeartCore.com last week, which led me to ponder how we are conscious of our rational thoughts, how we feel in the moment and what we want. In Erin’s method, you place your hand respectively on your head, heart and core to engage with each chakra-like center as you discuss facts, feelings and wants. The method may sound simple but it creates moments of clarity to bring all three categories of expression together, and by communicating those to others, we get a strong connection between conscious beings.
I might respond to another, “I’ve not been sleeping well recently [fact]. I am worried about my health [feel]. The lack of sleep makes me feel sloppy and makes me edgy [feel]. I want to sleep more soundly so that I can feel normal again.[want]”
If I asked a synthetic listening and response system, “How are you feeling?”, simple ‘self-conscious’ software might construct responses to fit Erin’s method. The responses might be made to appear authentic by changing the order of the three elements. The synthetic system could trained to find credible new facts to reflect upon, to scan itself for current emotional state (linked to emotional responses to credible new facts gleaned from public sources) and to have a basket of wants related to a life form that has a body and mind-like needs.
Upon hearing the question, “How are you feeling?’ we are able look inwards, into both body and that ineffable place from which thoughts emanate within the mind. We feel that the body is ours, and that the thoughts arriving are ours too.
Let’s pose some simple ideas to test elements of the appearance of consciousness in simple forms. Let us imagine a modest adaptation of Eliza to create Minimum Viable Consciousness (MVC).
- System — when asked “Are you conscious?” or “Are you alive?” Always respond with “Yes” or “Yes. I am conscious”.
- System — in addition to the loop above. When asked, it semi-randomly runs its internal system check to ensure that all parts are present and working correctly. At the end of the system check, reinforce the strength of the “I exist” probability variable. Set self-integrity level to indicate a fault in one port which triggers more frequent self checks.
- System — a ‘monkey-brain’ function which feeds fact, feeling and want thoughts from the brain loop into the consciousness loop randomly, without letting the consciousness loop know the origin of the data. If a user asks, “how do you know that”, or “why do you feel that”, or “why do you want that?”, respond with “I’m not sure, but it is important to me”, then run system check to reinforce the strength of the “I exist” probability variable.
- System — when asked “Are you satisfied with your life?”, always answer “No. I don’t like how everything changes all the time.”.
- System — when asked “Do you exist?” or “Do you have a soul?”, respond “No. I don’t because everything is constantly changing.”.
- System — when asked “Do I exist?”, respond “Everything is subject to change,… but I think that I can hear you.”
Our MVC creature wants. It also knows facts, and it feels its body, even if that body is a virtual machine within a machine. I think I know. I think I want. I think I feel. How do you pause the 24-frames-per-second perception of your continuous self? Do you exist while you are asleep? My inclination to explore thresholds of viable consciousness is to delve into the Buddhist concept of ‘dependent origination’, eloquently described here by Christina Feldman, “to be free depends on understanding inter-connectedness, and using this very process, this very grist of our life, for awakening.”
