From LLMs to Omniscient AI Machines: Navigating World's Complexity and Predicting the Unpredictable
"Omniscient AI Machines" is an extension of "Global Science and Engineering (GSE): Omniscient AI Technology: Ontological Engineering for Ontological Machines: Towards Ideal Machinery" and "Future AI: Real AI & ML & DL: Machine Ontology + Science + AI + ML + LLMs +..."
The idea of all-knowing machines and computer programs comes from science fiction, and now it is under study, development and deployment.
The World Wide Web was the start of that collection of world knowledge, enabling the large language foundation DL models like Galactica, ChatGPT, LaMDA, GPT-4, or Gemini having broad general knowledge and problem solving abilities.
The LLMs life cycle is encompassing phases like problem definition
But such omniscient systems do not have any understanding of what they are producing, any communicative intent, any model of the world, or any ability to explain the truth of what they are saying, so likened to stochastic parrots.
For example, without any knowing or intelligence or understanding ChatGPT generates humanlike conversational dialogue trained with RL via human feedback and reward models ranking the best responses, responding to questions and compose various written content, including articles, social media posts, essays, code, emails, product descriptions, summaries of transcripts, meetings and podcasts, simple explanations of complex topics, law briefs, translations, jokes or memes, blog posts.
Real Machine Intelligence and Learning necessitates machine philosophy, science, technology, engineering and mathematics (STEM), simulating and modeling reality, not human intelligence, in computing machinery.
Or, Omniscient AI is NOT the simulation of human intelligence or human brain or human behavior or human business or human body in machines.
AI must be proactive and interactive, scientific and technological and exhibit comprehensive world knowledge and scientific learning and reasoning, predicting the predictable and humanly unpredictable.
From LLMs to Omniscient AI Technology
We argue how it is possible to create omniscient AI machines which are knowledgeable to predict the unpredictable, to know the unknowable to anticipate the unanticipated.
Omniscient AI Machines: World Learning and Inference and Interaction Machine: Reality/Causality/Science > Statistics/Data Science > Mathematics/Set Theory/Optimization/Calculus/Linear Algebra/Probability > Programming/Algorithms/ANI/ML/DL/Neural Networks > LLMs/GPT/ChatGPT/Gemini > Software/Hardware/Infrastructure > Real-World/Ontological AI Applications
The world is getting exponentially more complex, including nature, living organisms, our planet and the entire universe, and the way to effectively interact with it is to understand its unpredictable and chaotic processes, knowing the unknowable, predicting the unpredictable.
An unanticipated and unpredictable change, process, disorder or “black swan†event can impact without warning and wreak horrific destruction and death, whether it’s the planetary pandemics, catastrophic quakes, stock market crushes, political revolutions, civil wars, military conflicts, local, regional or international, or, especially. advanced human-like AI systems.
For such a human-simulating AI & ML technology "takes our jobs, reduces our wages, increases inequality, threatens our health, ruins the environment, degrades our society, corrupts our children, impairs our humanity, threatens our future, and is ever on the verge of ruining everything", as a black swan event for our techno-optimists [The Techno-Optimist Manifesto, Marc Andreessen]
However, black swan events recognized in diverse fields, including economy and finance, politics and history, science and technology are not foreseeable by the usual calculations of correlations, regression, standard deviation, or reliability estimation, and probability predictions from big data analytics or the statistical AI of ML, DL, and ANNs.
We can never predict the future and emergent disruptions—but we can prepare for unpredictability and uncertainty by having the real AI-based models with all possible and impossible what-if scenario in silico.
Real AI as Machine's Predicting the Unpredictable could expect the unexpected, anticipate the anticipated, forecasting big effects from small causes, determinism from chaos, patterns from big data, identifying the time, location, and magnitude of future disruptions.
Real intelligence vs. simulated human intelligence is the power to process, organize?and leverage an infinite mass of world's data into knowledge?and intelligence as insights and solutions, algorithms and models, discoveries?or explanations,?predictions?and externalities, intelligent actions and interactions.
The real AI's world is modelled and simulated as a universal multi-entity, multi-network, dynamic network diagram, where a meta-network is a multi-mode, multi-link, multi-level network of entity variables with all possible interactions.
Multi-mode means that there are many types of nodes; e.g., variables, data elements, nodes people and locations.
Multi-link means that there are many types of links. Multi-level means that some nodes are members of other nodes, such as a nested network composed of people and organizations.
The world's complexity, from a cosmological complexity to computational complexity, in all its possible forms and kinds, complex systems and phenomena, from organisms and the human brain and social systems to a global climate and ecosystems to the entire universe, is modelled, simulated by real, true or scientific AI.
Real AI for the World's Complexity
In the complex dynamic world of nature or universe or reality, all is interrelated, everything is interconnected, nothing exists out of causal chains and circles and dynamic networks, complicated situations, processes or intricate interactions.
Such a universe of world's complexities has numerous elements, parts, components or data units and numerous forms of associations, correlations or relationships among the elements interacting in multiple ways, creating the patterns of interactions ("what/who interacts with what/who"):
RAI: The World of Complexities: Causal Interaction Networks {Entities and Interactions, Complex Systems, Topologies and Metrics, degrees, paths, centralities, transitivity, chains, feedback loops, etc.} + Scientific & Mathematical & Statistical & Computational Modeling [Applied Mathematics, Science, Network Science, Graph Theory, Network Theory, Dynamic Network Analysis] + ANNs [AI/ML/LL models]
In the world of nature and human societies, all substantial phenomena have causal interrelationships
The interaction networks with causal cycles of matter, energy and Information, life, intelligence and technology, involving the interactions of physical, geological, chemical, biological, ecological, social and technological processes, is the aim of world knowledge ranging from?natural sciences?to?cognitive science to social sciences?to?technological sciences to formal sciences, with all possible multi-disciplinary interactions.
Network science of complex networks is "the study of network representations of physical, biological, and social phenomena leading to predictive models of these phenomena."
Its Network theory as "the study of the way elements in a network interact" has applications in many disciplines, including?statistical physics,?particle physics, computer science,?electrical engineering,?biology,?archaeology,?linguistics,?economics,?finance,?operations research,?climatology,?ecology,?public health,?sociology,?psychology,?neuroscience, AI, ML, and DL.
Applications of network theory include?computer networks, the?World Wide Web,?Internet,?gene regulatory networks, metabolic networks,?neural networks, social networks,?epistemological?networks, semantic networks, logistical?networks, transportation networks, as well as Bayesian networks or Artificial Neural Networks of special topologies:
All the complexities of our nature and life marked with non-linearity, feedbacks, circularities, randomness, collective dynamics, self-organization, chaos, entropy, hierarchy, emergence, uncertainty, statistics, probabilities, etc.
So, our actions always produce something unknown, as unintended effects or unanticipated outcomes, as side effects or externalities, which are unpredicted and unexpected, unplanned or unintended. It is due to the world's complexity as well as a lack of causal knowledge, partial models and theories, errors and biases and prejudices, ignorance, self-deceit or groupthink.
In social sciences, such a causal complexity law, rule or pattern goes as the law of unintended consequences (unanticipated consequences or unforeseen consequences, knock-on effects), "the outcomes of a purposeful action that are not intended or foreseen".
"Almost all?environmental problems, from chemical?pollution?to?global warming, are the unexpected consequences of the application of modern technologies.?Traffic congestion, deaths and injuries from car accidents,?air pollution, and global warming are unintended consequences
Or, nobody of superpowers could predict that nuclear weapon would prevent nuclear war, as far as our world has been existing under the "mutual assured destruction", "MAD", military doctrine, when "each side has enough nuclear weaponry to destroy the other side. Either side, if attacked for any reason by the other, would retaliate with equal or greater force. The expected result is an immediate, irreversible escalation of hostilities resulting in both combatants' mutual, total, and assured destruction".
Now, it is clear that Tesla's "superweapon that would put an end to all war" is nothing but a non-human, real superintelligence (RSI), and the global AI arms race is the road to nowhere or our extinction.
Predicting the Unpredictable
Real intelligence vs. simulated human intelligence is the power to process and organize?an infinite mass of world's data into knowledge?and intelligence as discoveries?or explanations,?predictions?and prescriptions about the world, its global and local behaviors.
First of all, Real AI predicts the unpredictable, unknown and unexpected, what goes for human intelligence as unanticipated consequences, side effects or externalities, and together with identifying patterns, detecting trends and making exact predictions about future outcomes, as expected, planned or calculated effects.
To create the digital clones of humans, or the human body/brain/brains/behavior/business, is to create the computational replicas of human beings, with all the unexpected consequences.
Human-like AI has a limited understanding of the world, and its LL models can only predict what they were trained on. If the web data used to train prediction/ML models is biased, the predictions it makes will also be biased: GIGO.
Still, many human AI stakeholders and experts believe in human-level and human-like AI, unknowingly creating the global AI/AGI/ASI technology arms race to human extinction.
How Unpredictable could be Digital Technologies?
Again, most digital technologies have both positive and negative outcomes on humanity and society, especially digital AI technologies.
领英推è
Most modern technologies have negative consequences that are unplanned and unpredictable, but avoidable.
All stakeholders and organizational parties have to strategically consider how they will eliminate or minimize negative outcomes emerging from the global AI technology arms race.
In the short consensus paper, Managing AI Risks in an Era of Rapid Progress, it was outlined risks from upcoming, advanced AI systems, as large-scale social harms and malicious uses, as well as an irreversible loss of human control over autonomous AI systems.
Meantime, the European Parliament has adopted the world's first "Artificial Intelligence Act: deal on comprehensive rules for trustworthy AI", It features:
- Safeguards agreed on general purpose artificial intelligence
- Limitation for the of use biometric identification systems by law enforcement
- Bans on social scoring and AI used to manipulate or exploit user vulnerabilities
- Right of consumers to launch complaints and receive meaningful explanations
- Fines ranging from 35 million euro or 7% of global turnover to 7.5 million or 1.5% of turnover
Our post explains why the human-like, simulated, unreal or unrealistic AI/AGI/ASI, a common topic in science fiction and futures studies, is existentially risky, and why the real/true/scientific AI/AGI/ASI is the intelligent and sustainable future.
We make plain and clear why the Real AI Technology should replace the human AGI technology; for the race to AGI/ASI, as a human-like and human-level AI between great powers is "A Race to Extinction: How Great Power Competition Is Making Artificial Intelligence Existentially Dangerous".
Unlike the human-imitating-replacing AI/AGI, the Real AI is the non-human complement of human intelligence, individual or collective
True, Real or Genuine AI is requesting for its creation world knowledge and intelligence, including STEM and Art & Design, Business, Health & Medicine, Humanities & Social Science, as well as computing/data/information science and engineering, cybernetics, dynamic networks analysis, systems engineering and other multi-, meta-, trans-disciplinary knowledge, structured by global formal ontology, mathematics, causal logic, semiotics and NL.
It is argued that there is no real and true AI in existence, yet, in any ways and forms. It is due to the fact that the global generative AI arms race is overwhelmed with self-delusion and commercial fraud, overhype and buzzing and inflated expectations.
There are commercial simulated AI/ML models, as LLMs, GPT-x or OpenAI's ChatGPT, Google's Gemini, where LLMs are rather stochastic machine rote learning calculators for numerical strings/tokens, as text, images, video, audio, and code data, without ZERO learning and reasoning, understanding or intelligence.
Omniscient Machines as Global Predictors
To navigating the world's complexity and predicting the unpredictable, it must be some all-knowing AI platform powerful to predict causal timelines of events, human life, societies, nature and the universe.
Today, future is completely unpredictable, we are even unaware of the pandemics to come.
On a small practical scale, we have predictive analytics with its machine deep learning models, exploiting pattern recognition, to analyze current and historical facts to make predictions about future.
As parts of predictive techniques, there are a lot of forecasting methods, qualitative and quantitative, added up with strategic foresight or futures [futures studies, futures research or futurology] with no big utility. It all revolves around pattern-based understanding of past and present, to explore the possibility of future events and trends.
On a large scale, the nature timeline is a big mystery even afterwards, not mentioning the universe timeline, where we have poor ideas about its history, far future and ultimate fate.
Resources
AI Superintelligence/AI/AGI/ASI/Trans-AI
Real AI Project Confidential Report: How to Engineer Man-Machine Superintelligence 2025: AI for Everything and Everyone (AI4EE); 179 pages, EIS LTD, EU, Russia, 2021 (see Supplement).
[Mind, Machine Intelligence, AI, Artificial, Automated, Autonomous, Automatic Intelligence, Machine Learning, Deep Learning]: exposing the great unknown unknowns. Part III.
SUPPLEMENT
Real AI Project Confidential Report: How to Engineer Man-Machine Superintelligence 2025: AI for Everything and Everyone (AI4EE); 179 pages, EIS LTD, EU, Russia, 2021
Content
The World of Reality, Causality and Real AI: Exposing the great unknown unknowns
Transforming a World of Data into a World of Intelligence
WorldNet: World Data Reference System: Global Data Platform
Universal Data Typology: the Standard Data Framework
The World-Data modeling: the Universe of Entity Variables
Global AI & ML disruptive investment projects
USECS, Universal Standard Entity Classification SYSTEM:
The WORLD.Schema, World Entities Global REFERENCE
GLOBAL ENTITY SEARCH SYSTEM: GESS
References
Supplement I: AI/ML/DL/CS/DS Knowledge Base
Supplement II: I-World
Supplement III: International and National AI Strategies