Artificial general intelligence
Artificial general intelligence is a hypothetical form of machine intelligence capable of learning and performing a broad range of intellectual tasks across domains with human-like generality or greater.
Introduction to Generative AI, Second Edition
Artificial general intelligence is a hypothetical AI capability encompassing broad learning, reasoning, and adaptation across tasks at roughly human or greater generality.
- How it is used
- Frames debate about whether broad AI capabilities amount to general intelligence and how that debate shapes public and corporate priorities.
- How deep it goes
- Compares competing definitions and concrete claims but stops short of a settled, testable criterion for AGI.
- Examples
- Uses a vehicle choosing an alternate route and claims about GPT-4’s cross-domain performance to distinguish evidence from interpretation.
- What you can do after
- Evaluate/understand AGI claims by checking their definition and distinguishing broad performance from proof of general intelligence.
Read how Introduction to Generative AI, Second Edition explains artificial general intelligence →
Sutskever's List
The notion of an artificial system possessing general, human-like intelligence rather than merely performing limited tasks.
- How it is used
- Frames human-like general problem-solving as a research goal distinct from narrow task performance.
- How deep it goes
- Defines the goal and traces debates over its meaning, but offers no definitive test or construction method.
- Examples
- Uses OpenAI’s early research milestones and opposing interpretations of GPT-4’s cross-task performance.
- What you can do after
- Evaluate/understand claims that broad model capabilities demonstrate progress toward AGI without treating those claims as settled.
Read how Sutskever's List explains artificial general intelligence →
WordPress Power Toolkit: Harness AI to build next-level websites
Artificial general intelligence is AI capable of performing a broad range of intellectual tasks at or above human level.
- How it is used
- Distinguishes broadly capable AI from specialized systems when comparing AI tools and their limits.
- How deep it goes
- Defines AGI through broad human-level intellectual performance and contrasts it with task-specific AI without detailing technical mechanisms.
- Examples
- Uses disease diagnosis, chess, and Go as examples of narrow AI exceeding humans in specific tasks.
- What you can do after
- Evaluate/understand whether an AI claim describes specialized performance or broad general intelligence.
Deep Learning with R, Third Edition
General intelligence is broad intellectual capability that enables a system to perform a wide range of cognitive tasks rather than a narrowly defined task.
- How it is used
- Frames artificial general intelligence as the broader target beyond systems optimized for narrowly defined tasks.
- How deep it goes
- Defines AGI through adaptation, uncertainty, and novelty, but does not provide an implementation or construction method.
- Examples
- Uses Deep Blue, chess, Go, MNIST, ImageNet, video games, and fixed-rule systems to contrast narrow skill with generalization.
- What you can do after
- Evaluate AGI claims by distinguishing adaptation to novel tasks from high performance achieved through task-specific data or rules.
Read how Deep Learning with R, Third Edition explains artificial general intelligence →
Related concepts
- narrow aicontrasts with · Narrow AI is an artificial-intelligence system specialized for a limited task or domain rather than general-purpose intelligence across unrelated domains.
- superintelligencecontrasts with · Superintelligence is intelligence whose general cognitive abilities substantially exceed human capabilities across many or most cognitive domains.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.