Artificial General Intelligence (AGI) refers to artificial intelligence that exhibits cognitive abilities comparable to humans across a wide range of domains. This is in contrast so-called "narrow" AI that may exhibit human-level capabilities, but only at specific tasks or in highly controlled environments.

Expert's definitions about what exactly constitutes AGI differ, but many argue that AGI should learn from its environment, reason about diverse topics (including natural language, math and logic, the spatial world around us, etc.), and apply knowledge to solve complex problems in novel situations –just like a human would. AGI is (as of the writing of this encyclopedia entry) still largely the realm of science fiction, and achieving it is the "holy grail" of data science research.

How is Artificial General Intelligence Different from Other Forms of AI?

While "narrow" AI is designed to perform specific tasks that have limited scope. AGI would be able to generalize from one task to another and would be able to function autonomously, without the need for human intervention. It would be able to learn and adapt over time and would be capable of making its own decisions and judgments.

Are There Existing Examples of Artificial General Intelligence?

Despite what you may have read on Twitter –No, there are not. AGI is still largely a theoretical concept and there are no known examples of true AGI systems in use today. Researchers, entrepreneurs, politicians, and writers continue to theorize how AGI might arise and about how it might help or harm humanity.

Hey Siri, What Do You Think of My New Boyfriend?

Despite this, there are AI systems that exhibit some of the characteristics of AGI. For example, the Large Language Models (LLMs) have been an impressive breakthrough at understanding human language and responding conversationally. These systems however do not meet the definition of AGI –even though they do generate intelligent human-like text and can perform a much wider range of tasks than their predecessor algorithms.

Other areas of advanced AI systems that exhibit some characteristics of AGI include robotics, computer vision, and reinforcement learning systems (think Google DeepMind) –which are capable of physical manipulation and responding to their environments. Despite these examples, true AGI remains a long-term goal.