Artificial General Intelligence (AGI): From History to the Future

ORCID: 0009-0001-2652-7017The paper presents an interdisciplinary overview of artificial general intelligence (AGI) – from its historical roots to future implications – with an emphasis on philosophical, ethical, and societal challenges. It starts from an analysis of the evolutionary path of artificial intelligence, from early symbolic systems and "AI winters" to today's dominance of deep learning and large language models (LLMs), highlighting the fundamental differences between narrow (ANI) and general intelligence.The first part defines the key characteristics of AGI – abstract reasoning, common sense, transfer learning, and metacognition – along with a critical analysis of why the "common sense problem" is the greatest obstacle to machine intelligence. The second part examines in detail the potential benefits of AGI (solving global problems, scientific acceleration, economic efficiency) alongside an analysis of existential risks: the alignment problem, instrumental convergence, mass unemployment, geopolitical arms races, and the erosion of privacy.A special contribution is the analysis of the "post‑purpose economy" – a scenario in which AGI automates all cognitive tasks, creating a "useless class" of people without an economic and social role. The paper examines empirical evidence from UBI experiments around the world (Finland, Kenya, Canada, Spain) as a potential mechanism for wealth redistribution, but points to its limitations in addressing the existential void.In the technical section, the paper reviews contemporary approaches to the safe development of AGI: Oracle AI, inverse reinforcement learning (IRL), Constitutional AI, mechanistic interpretability, and global coordination. It specifically analyzes arguments for and against the possibility of creating AGI, including Searle's "Chinese Room," Penrose's quantum arguments, embodiment, and materialist approaches, with reference to the role of quantum computing and neuromorphic hardware.The paper concludes that the development of AGI is not merely a technical challenge, but above all a philosophical, ethical, and anthropological one that requires global cooperation, an interdisciplinary approach, and a deep re‑examination of human values. Without conscious political and moral intervention, the AGI revolution risks being the greatest increase in inequality in human history, rather than a universal accelerator of progress.In its final part, the paper announces the author's book "Theory of the Metasystem: Toward a Model of the Conscious Machine," which presents an original philosophical framework according to which consciousness can be an emergent property of controlled conflict between the analytical and intuitive principles of information processing, governed by a third principle that sustains creative tension. This theory offers an experimental framework for testing hypotheses about machine consciousness, shifting the focus from the question "can a machine be conscious" to "under what conditions can consciousness emerge."Keywords: artificial general intelligence, AGI, alignment problem, artificial consciousness, AI ethics, reinforcement learning, UBI, existential risk, Metasystem Theory, philosophy of mindType of paper: review article with original philosophical contributionNumber of references: 32