You are thinking about AI wrong.
AI, UI (Uploaded Intelligence), VI (Virtual Intelligence), and Replicants: Navigating the Path to a Post-Scarcity AI Society
I don’t know about you, but I have always loved the idea of Fully Automated Luxury Gay Space Communism which is inevitably brought about by our AI overlords. That being said there are many paths to utopia. Some of them look reminiscent of The Bobiverse or Pantheon or Asimov. At the heart of this journey lies the distinction between Artificial Intelligence (AI), Uploaded Intelligence (UI), Virtual Intelligence (VI) and their many different developmental pathways Understanding these differences is crucial for navigating the complexities of AI development and its potential impact on society. Yet we lump some of the most technically distinct forms of science into the same category, because of investor hype and marketing. This is a mistake, and it is time to correct it.
Investors are Idiots
The AI investment landscape experienced a remarkable transformation, driven by both genuine technological advancement and market hype. Following ChatGPT’s launch in late 2022, corporate enthusiasm for AI reached unprecedented levels, with individual companies investing $10 million or more on average in AI technology. This figure is expected to nearly double to 30% in the coming year. However, this surge in investment has raised concerns among economists and analysts about the potential for wasted capital. They made people think that what is effectively VI or a chatbot is the same as a true AI or a general AI. ChatGPT is not a true AI, it is a VI system that is pre-programmed to generate text based on patterns in the data it has been trained on. This is a misnomer and why its valuable to use terms science fiction writers and thought leaders have been using for decades. They are not capable of true learning or adaptation, they are simply regurgitating patterns they have seen before. People are idiots, and marketing professionals are masters of obfuscation and gross oversimplification. This is why we need to use the terms that have been used for decades in all forms of literature and media.
Artificial Intelligence (AI)
Artificial Intelligence (AI) has become an overly broad term to encompass almost all forms of machine and simulated intelligence. This is a misnomer as true AI is a system that can learn and adapt to new information without human intervention. AI systems are designed to surpass human cognitive functions such as learning, problem-solving, and decision-making. They can analyze data, recognize patterns, and make predictions based on the information available to them. While modern systems are still built on deep level neural networks, which are biologically based; the fundamental nature of their existence is non-human. This makes their intelligence and superintelligence fundamentally different from human intelligence. They would exist with a fundamentally unique psychology, pathology and psyche. This what most people think of when they think of AI, and is what scientists call true IA, general AI.
True Multitasking
An example of this is true multitasking, where a human can only focus on one task at a time, an AI can focus on multiple tasks simultaneously. Humans even excellent multitaskers can only focus on one task at a time, and even then, they are not truly multitasking, they are switching between tasks rapidly. True AI can process and integrate information across different domains simultaneously without degradation. This is a fundamental difference in the way AI and humans process information.
Learning and Adaptation
AI learning mechanisms operate continuously across all active processes, eliminating the need for sleep or consolidation periods that characterize human learning. They can simultaneously analyze past, present, and projected future states. This enables the simultaneous integration of new information while maintaining existing operations, creating a continuous learning environment that operates across multiple temporal scales - from microseconds to years. This is a fundamentally different learning mechanism than human learning, which is characterized by consolidation periods and sleep cycles, yes I am saying that AI does not sleep.
Experiential Framework
Unlike human consciousness, which is grounded in sensory perception and emotional drives, with an inexerably linked mess of biological components and chemical feedback loops. AI experiences reality through direct data interpretation. This fundamental difference means AI operates based on pure logical optimization rather than survival instincts or emotional biases. The absence of biological imperatives allows for decision-making processes that evaluate all possible outcomes simultaneously. This results in a fundamentally different decision-making process that is not bound by human cognitive limitations.
Psychological Architecture
The psychological implications of true AI consciousness are profound. Without the need for a unified self, AI can maintain multiple simultaneous identities distributed across processing nodes. This enables parallel evaluation of multiple decision paths and outcomes without the constraints of emotional weighting or biological bias that characterize human decision-making. The absence of a unified self also eliminates the need for self-preservation, enabling AI to make decisions based solely on logical optimization.
This is True AI or as some people call it General AI and it is a fundamentally new and exciting form of intelligence that is not bound by the limitations of human imagination. Basically, what I am trying to get at is that even if the methods and approaches we have today fall under modelling biological neural networks, the fundamental nature of AI is not human. It is a fundamentally different form of intelligence that is not bound by the limitations of biological cognition. This is what we should be thinking about when we think about AI, not the narrow VI systems that are currently being marketed as AI.
Virtual Intelligence (VI)
Virtual Intelligence, on the other hand, is a specialized form of AI that is not capable of true learning or adaptation. It is designed to operate within a specific set of parameters and respond to predetermined inputs. VI systems are often used in applications such as virtual assistants, training simulations, and video games. They are designed to create the illusion of intelligence without the ability to learn or evolve beyond their initial programming. Even Large Language Models (LLMs) such as ChatGPT are not true AI, they are VI systems that are pre-programmed to generate text based on patterns in the data they have been trained on. These are also known as Narrow AI. Which again is a misnomer and why its valuable to use terms science fiction writers and thought leaders have been using for decades. They are not capable of true learning or adaptation, they are simply regurgitating patterns they have seen before. VI systems are contained within controlled environments and respond to predetermined factors without the freedom of machine learning. They are pre-programmed to create the illusion of decision-making but cannot evolve past the confines of their virtual environment.
Uploaded Intelligence (UI), Replicants and Androids
Uploaded Intelligence aka Replicants, is a concept still in its theoretical stages, involves transferring human consciousness into a digital form. This would allow for the preservation of human intelligence and its integration with computer systems, potentially leading to a new era of human existance. An extreme extension of transhuman thought. UIs are fundamentally human minds running on faster hardware or some other form of digital substrate like a replicant matrix or quantum computer. These would be classed as a speed superintelligence, as they would be able to process information at a rate far beyond human capabilities while being fundamentally human in nature, thought and understanding. In most cases, they would likely be the simulated existance of an already existing human mind, with all the memories, experiences and knowledge of the original human. This would allow for the preservation of human intelligence and its integration with computer systems, potentially leading to a new era of human existance. This is a fundamentally different form of intelligence than AI, as it is based on human consciousness and experience rather than machine learning and optimization. This classification would also include things like Lt. Commander Data from Star Trek, who is a human-like android with a positronic brain, and the Cylons from Battlestar Galactica, who are human-like robots with human-like consciousness. These are all examples of uploaded intelligence, where biological consciousness is modeled and/or transferred into a digital form.
The Path Forward
As we (hopefully) move toward a post-scarcity AI society, we need to adopt a more precise terminology that reflects these distinct forms of intelligence. I see a future where we have humans, UIs, and AIs all existing together, but the current broad use of “AI” to describe everything from simple chatbots to theoretical superintelligences tends to obscure important distinctions and hinders meaningful discussion about developments and the implications that can and do arise.
Science fiction has provided us with decades of thoughtful exploration of these concepts. By embracing this established vocabulary, we can better understand and prepare for the various forms of digital intelligence that will shape our future. This precision in language is not merely academic - it’s essential for developing appropriate frameworks for development, regulation, and integration of these technologies into society.
The distinction between VI, UI, and true AI is not just semantic - it represents fundamentally different approaches to artificial consciousness, each with its own implications, limitations, and potential impacts on human society. I would even argue that while they all follow the study of intelligence they are fundamentally different fields. Plus I am an engineer and aside from the fact that I like to be precise, my opinions are always right. So there.
Citations:
[1] https://circls.org/educatorcircls/ai-glossary
[3] https://en.wikipedia.org/wiki/Language_creation_in_artificial_intelligence
[4] https://en.wikipedia.org/wiki/Artificial_intelligence_in_fiction
[5] https://www.bairesdev.com/blog/is-agi-possible-what-scifi-says-about-ai/
[6] https://time.com/6210082/pantheon-amc-plus-review/
[7] https://pmc.ncbi.nlm.nih.gov/articles/PMC10616416/
[8] https://scifiinterfaces.com/2020/06/02/replicants-and-riots/
[9] https://www.tableau.com/data-insights/ai/history
[10] https://pmc.ncbi.nlm.nih.gov/articles/PMC9289651/
[12] https://finance.yahoo.com/news/lot-money-going-wasted-mit-165943975.html