Artificial general intelligence
Artificial general intelligence
(also AGI)
Artificial general intelligence definition
Artificial general intelligence (AGI) is a hypothetical form of artificial intelligence that can understand, learn, and apply knowledge across a wide range of tasks at a level comparable to humans or beyond.
See also: artificial intelligence, generative AI, machine learning
Characteristics of artificial general intelligence
- Learning and adaptability. Unlike narrow AI, designed for a specific task, AGI could learn and adapt to various tasks, much like a human.
- Reasoning. AGI could solve unfamiliar problems using logic and reasoning.
- Self-awareness. An advanced characteristic of AGI would be self-awareness or consciousness, though this remains a topic of debate.
- Transfer learning. AGI could use knowledge from one area in another, different context.
- Autonomy. AGI could work without human intervention, making decisions based on its learning and experiences.
- Emotional understanding. While still theoretical, true AGI might be able to understand and even replicate human emotions and nuances.
History of artificial general intelligence
- Early Philosophical Foundations. Philosophers like Plato and Descartes mused about the nature of human thought and the possibility of replicating it. They laid the groundwork for the later exploration of intelligence, both human and artificial.
- 19th century. Innovations like Charles Babbage’s analytical engine showed the earliest signs of programmable logic.
- 1930s-1940s. Alan Turing formulated the concept of a universal machine that could simulate any computable sequence. His 1950 paper, “Computing Machinery and Intelligence,” introduced the Turing test for machine intelligence.
- Dartmouth workshop (1956). The term “artificial intelligence” was first used and the goal of building AGI was first proposed.
- 1960s-1970s. This was a time of significant optimism about the potential of AI. However, technical and financial challenges led to periods known as “AI winters,” where funding and interest in the field shrunk.
- 1980s-1990s. As broader AGI goals remained elusive, the focus shifted to narrow AI, specialized for particular tasks.
- 21st century. A renewed interest in AGI arose with advancements in deep learning, neural networks, and computational power. Companies like DeepMind and OpenAI, among others, explicitly stated their goal of achieving AGI.
- Current state. The pursuit of AGI continues, with research focusing on versatile learning and wider task adaptability.