Levels of AI-Human Integration
Epistemic Status: common knowledge, just not to me.
The below are suggested levels of integration, when viewed from physical and ontological dimensions. The levels can be viewed in as a kind 2D "shock levels" of integration.
Introduction
When talking about human-AI integration, it may be convenient to use the analogy of the integration of computer systems, whereby the integration happens at different levels of communication (I/O) protocols, where there is physical layer at the bottom, and applications layer at the top.
Human minds can be thought of as knowledge hypergraphs, that interpret certain signals (external and imagined) as positive or negative stimuli (pain and reward), which reflects the trajectory of approach towards the internalized goals.
An ultimate example of integration can be thought of as the sense of single self, which is born when the communication is continual and complementary (for example, left brain hemisphere and right brain hemisphere continually exchange information via corpus callosum, and a human in most cases feels like a single person, though, the rare dissociative personality disorder does exist, even in a relatively well-connected minds).
Physical Levels:
- level: (symbiotic-level) when systems happen to live together, only exchanging the products of their metabolism. For example, Toxoplasma Gondii may be said to be integrated with a house cats (or a human), as it affects their behavior.
- level: (actuatory-level) when systems interact through exchange of physical force to perform the I/O. (e.g., typing).
- level: (sensory-level) when systems interact through direct stimulation of specialized sensors or sensory receptors to exchange information (e.g., direct stimulation of retina, and direct generation of photos in retina as a result of projecting thoughts as imagined matrices.).
- level: (inductive-level) when systems communicate through parallel non-sensory structures (such as parallel artificial neural network), with mutually inductive excitatory exchange.
- level: (reflective-level) when systems communicate through scanning the reflections of states of another system to determine what that system is trying to say. (For example, doing the red-light brain scan to understand what brain is trying to say, meanwhile giving interpreted feedback via letting the person see the internal computer state changes in raw form of visual electric charge distribution view.)
- level (mixed) when systems communicate via combinations of 1-5 levels.
Based on this physical level, modes of integrative communication can be devised. For example, we may not be very satisfied with seeing raw states of computer bits and electric charges, or raw data objects, but instead, may want o see the representation of those states in form of objects that our visual cortex is trained to recognize (such as pictures, texts). Vice versa, when viewing another person's mind, we may not want to view the raw representation of the scans, but instead, have them translated into imagery, and ultimately, equivalent sensory perceptions of another system. If the system is non-human, the perceptions may be desired to be translated into comparable emotional triggers about gain/loss with respect to goals of comparable personal significance to the systems.
Ontological Levels:
- level (emotion-only) a mind produces an emotion that represents the needs, and receives input back as as direct emotion triggers to comprehend whether or not the action they took is leading towards desired outcomes. For example, the relation of actions to economical indicators may be inferred from an economic model, and a an AI may have a direct trigger to the emotions to provide feedback to the mind to associate the collective risk landscape with personal emotions. (e.g., feel the physical pain because of prediction of long-term climate change).
- level (concept-only) a mind produces a program based on symbols, which then is parsed by another system in form of code review. The systems collectively give feedback about the code to be executed, and execute when reach some agreement.
- level (one a mutual ontological mapping and reward signals learning, with respect to personal goals of each one, and an exchange of complex signals with respect to negotiated common goals). A translation AI is employed (or a part of resources is dedicated) to learn the two or more minds, and provide a real-time exchange mapping
In the level of P6 and O3, we may see the emergence of a single new integrated mind. It might be, that this can only happen, when the goals with respect to the kind of life one wants to lead, are sufficiently similar, that the common goals can be negotiated, and be made a top priority. If that is not the case, we may see the "dissociative-personality disorder" type of "integration," which may be the current state of the world due to lack of capabilities to negotiate or shared data streams at some of the levels.
Further Work
Explore, how do the levels of AI-Human integration inform the rise of collective consciousness from the perspective of integrated information theory.