A learning path ready to make your own.

What is general artificial intelligence?

What is General Artificial Intelligence (AGI)? Executive summary: AGI (also called strong AI) denotes machines that can understand, learn, and apply intelligence across a wide variety of tasks and domains at a level comparable to—or exceeding—competent humans. Unlike narrow AI, which targets specific tasks, AGI requires flexibility, transfer, abstraction, reasoning, and efficient learning in novel situations. The topic spans technical, cognitive, philosophical, and policy dimensions. Definitions & core concepts AGI / Strong AI: Systems able to perform any intellectual task a human can, with comparable breadth and adaptability. Narrow / Specialized AI: Task-specific systems that may exceed humans on narrow tasks but lack generality. Generalist / Broad AI: Systems handling many tasks (multimodal, multitask) but not necessarily human-level across all domains. Key AGI capabilities: cross-domain transfer, abstraction, reasoning, planning, meta-learning, continual learning, commonsense, multimodal integration, and data-efficient learning. Historical overview (concise) 1950s–1980s: Symbolic AI, Turing Test, early cognitive architectures (SOAR, ACT-R). 1980s–1990s: Rise of connectionism and statistical learning; AI winters from unmet expectations. 2000s–2010s: Machine learning and deep learning deliver narrow superhuman systems in vision, speech, and games. 2018–2024: Foundation models (large transformers), multimodal and multi-task agents show broad capabilities but lack many core AGI features. Theoretical foundations Formal ideals: Universal intelligence, Solomonoff induction, and AIXI provide normative but often incomputable benchmarks. Computational constraints: Practical intelligence must trade off generality vs. time, memory, and data. Learning paradigms: Supervised, self-/unsupervised, reinforcement, meta-learning, continual learning, and causal/structured representation learning. Philosophical roots: Symbolic vs connectionist debate, embodied cognition, and cognitive architectures guide design choices. Key capabilities commonly required for AGI Broad generalization and transfer across domains. Efficient learning and meta-learning (few-shot adaptation). Abstraction, symbolic and causal reasoning. Robust planning and long-horizon reasoning under uncertainty. Continual learning and durable memory (episodic/semantic). Multi-modal perception and coherent integration. Social, language, and commonsense understanding. Creativity and invention in novel domains. Physical interaction and manipulation for embodied agents. Self-monitoring, reflection, and goal adaptation. Architectural approaches & research paths Scaling foundation models: Larger transformers on massive data (LLMs, multimodal variants); strengths in breadth, limits in reasoning, grounding, and sample efficiency. Neurosymbolic / hybrid: Combine neural perception with symbolic reasoning for systematicity and interpretability. Modular / cognitive architectures: Interacting modules (perception, memory, planner) inspired by cognitive science. Meta- & few-shot learning: Learn priors or update rules for rapid adaptation to new tasks. Continual / lifelong learning: Methods to avoid catastrophic forgetting and accumulate knowledge. Embodied / developmental: Agents that learn via interaction and curricula in simulation or physical environments. Planning & world models: Model-based RL, learned simulators, and search for long-horizon decision-making. Causal / structured representation learning: Aim for causal, robust representations supporting interventions and transfer. Benchmarks & evaluation Behavioral tests: Turing Test, ARC, Animal-AI Testbed, BIG-bench and task suites probe reasoning and adaptability. Practical axes: breadth, efficiency (data/compute), robustness, transfer, and safety/alignment. Formal metrics (e.g., universal intelligence) are principled but often impractical. State of the art (as of 2024) LLMs (GPT, PaLM, LLaMA) show strong language skills, in-context learning, and some reasoning, but suffer hallucinations, brittleness, and grounding limits. Multimodal generalist agents (e.g., Gato) demonstrate multi-task behavior but remain far from AGI. Deep RL achieves domain-specific superhuman performance but with poor transfer and high sample complexity. Robotics and cognitive architectures make steady progress; general-purpose physical and cognitive agents are still unresolved challenges. Applications (near-term and hypothetical) Near-term: assistive agents for writing, coding, summarization; decision support; partial automation of knowledge work; robotics for constrained tasks. If AGI emerges: rapid scientific discovery, universal personal assistants, autonomous research/engineering agents, advanced policy and legal reasoning systems. Risks, safety, and alignment Technical risks: specification gaming, goal misgeneralization, distributional failures, instrumentality, and interpretability gaps. Societal risks: labor displacement, concentration of power, misuse (cyber, bio, influence), and governance lag. Existential concerns: long-term risks if superhuman AGI is misaligned; debates center on controllability and value alignment. Safety research directions: capability control, robust reward design, value alignment, scalable oversight, interpretability, and multi-agent dynamics. Governance, policy & timelines Policy needs: testing and certification standards, transparency, liability frameworks, export controls, and international coordination. Timelines: expert opinions vary widely; arrival depends on scaling, algorithmic breakthroughs, compute/data trends, and social constraints. Key uncertainties: whether scaling suffices, whether new conceptual breakthroughs are required, and how societal choices shape incentives. Ethical & philosophical considerations Questions about consciousness, moral status, and rights of advanced agents remain unresolved. Responsibility and liability: who is accountable for agent actions—developers, deployers, or operators? Value pluralism: whose values should AGI reflect, and how to reconcile cultural differences? Practical recommendations For researchers: prioritize safety, robustness, interpretability, human oversight, and hybrid approaches; benchmark generalization and sample efficiency. For policymakers: fund independent safety research, create risk assessment frameworks, and coordinate internationally on high-stakes capabilities. For public & institutions: foster broad dialogue, prepare workforce policies (reskilling, safety nets), and include diverse stakeholders in decision-making. Conclusion: AGI aims to achieve human-level flexible intelligence across domains. Recent progress in large-scale models and multimodal agents has advanced generalist capabilities, but core challenges remain—causal reasoning, sample-efficient lifelong learning, aligned goal-directed behavior, and safe deployment. Progress will likely be incremental and interdisciplinary, requiring parallel technical work and governance, ethical, and societal preparation. Selected further reading Stuart Russell & Peter Norvig, "Artificial Intelligence: A Modern Approach" Shane Legg & Marcus Hutter, "Universal Intelligence" (2007) Nick Bostrom, "Superintelligence" (2014) Richard Sutton, "The Bitter Lesson" (2019) Gary Marcus, "Rebooting AI" (2019) Recent technical papers from OpenAI, DeepMind, and related groups on foundation models, multimodal agents, RL, and safety.

Let the lesson walk with you.

Podcast

What is general artificial intelligence? podcast

0:00-3:11

Follow the trail that experts already trust.

Resources

Turn quick sparks into lasting recall.

Flashcards

What is general artificial intelligence? flashcards

16 cards

Question

Click to flip
Answer

Prove the idea before it slips away.

Quizzes

What is general artificial intelligence? quiz

12 questions

Which statement best captures the definition of Artificial General Intelligence (AGI) as described in the content?

Read deeper, connect wider, own the subject.

Deep Article

What is General Artificial Intelligence? =======================================

Executive summary


Artificial General Intelligence (AGI)—also called general artificial intelligence or strong AI—refers to a machine or system that can understand, learn, and apply intelligence across a wide variety of tasks and domains at a level comparable to (or exceeding) that of a competent human. Unlike narrow (or weak) AI, which is built for specific tasks (e.g., image recognition, language translation, game playing), AGI would display the flexibility, transfer, abstraction, reasoning, and learning capabilities necessary to handle novel situations it was not explicitly trained on.

This article provides a detailed, multidisciplinary survey of AGI: conceptual definitions, historical context, theoretical foundations, candidate architectures and research directions, evaluation methods, current state of practice (as of 2024), practical implications, safety and governance challenges, and likely trajectories and uncertainties.

Definitions and core concepts


  • Artificial General Intelligence (AGI): A system that can perform any intellectual task that a human being can, with comparable breadth, adaptability, and learning efficiency across domains.
  • Narrow AI / Specialized AI: Systems engineered for a single or small set of tasks (e.g., speech recognition, chess playing). They may exceed human performance on their tasks but lack generality.
  • Generalist AI / Broad AI: Often used to denote systems that handle many tasks (multimodal, multitask) but not necessarily all tasks at human-level generality.
  • Strong AI vs Weak AI: Philosophical division—strong AI refers to genuine understanding/mind (equivalent to AGI), weak AI refers to tools and simulators of intelligence.
  • Key capabilities associated with AGI: cross-domain transfer, abstraction, reasoning, planning, meta-learning, continual learning, common-sense understanding, flexible goal pursuit, robust interaction with the physical and social world, and efficient use of limited data.

Historical overview


  • 1950s: Foundational ideas emerge. Turing proposes the "imitation game" (Turing Test) as a behavioral criterion for intelligence. Early symbolic AI (Newell & Simon) and the cybernetics community explore formal reasoning and problem solving.
  • 1960s–1980s: Symbolic AI (knowledge representation, logic, planning) dominates. Work on cognitive architectures (e.g., SOAR, ACT-R) aims for broad cognitive coverage.
  • 1980s–1990s: Emergence of connectionist approaches, backpropagation, and statistical learning. The AI winter(s) reflect unmet expectations around general-purpose intelligence.
  • 2000s–2010s: Rise of machine learning, big data, and deep learning. Specialized systems achieve human-level or superhuman performance in narrow domains (vision, speech, games).
  • 2018–2024: Foundation models (large pre-trained transformer-based models) show broad capabilities across many tasks via fine-tuning and prompting. Multimodal and multi-task agents (e.g., Gato, multimodal transformers) demonstrate an incremental move toward generality but still lack many core AGI capabilities.
  • Present debates center on whether scaling current paradigms (bigger models, more compute/data) or combining approaches (symbolic + connectionist + embodied) is most likely to produce AGI.

Theoretical foundations


Formal characterizations

  • Universal intelligence (Legg & Hutter, 2007): Intelligence is the expected performance of an agent across a wide range of environments, weighted by their simplicity (Occam's razor). This provides an abstract, formal notion but is not computable in practice.
  • Solomonoff induction and AIXI: Theorize about optimal prediction and acting agents with access to all computable hypotheses; AIXI is a theoretically optimal but incomputable agent that integrates Solomonoff induction and sequential decision theory. These provide normative ideals but are not practical.
  • Computational complexity: Intelligence is constrained by computational resources — time, memory, and data. Practical AGI must navigate trade-offs between generality and tractability.

Learning paradigms relevant to AGI

  • Supervised learning: Learning from labeled examples—powerful but data-hungry and often narrow.
  • Unsupervised / self-supervised learning: Learning internal representations from raw data signals; foundational to modern large models.
  • Reinforcement learning (RL): Learning via trial-and-error with rewards; important for sequential decision-making and control.
  • Meta-learning (learning-to-learn): Systems that adapt rapidly to new tasks from few examples—key for data-efficient generalization.
  • Continual learning / lifelong learning: Learning across streams of tasks without catastrophic forgetting; necessary for accumulating knowledge over time.
  • Causal learning and structured representation learning: Learning cause-effect relations and interpretable world models to support robust planning and counterfactual reasoning.

Philosophical and cognitive foundations

  • Symbolic vs connectionist debate: Historically, symbolic AI emphasized explicit representations and logic-based reasoning; connectionist approaches (neural nets) emphasize distributed representations and learning from data. Hybrid/neurosymbolic approaches try to combine strengths of both.
  • Embodied cognition: Argues intelligence arises from interaction with the environment; embodiment (sensors, actuators) may be critical for grounding concepts and achieving robust generality.
  • Cognitive architectures: Models like SOAR and ACT-R aim to capture human cognitive processes; they provide frameworks for modular cognitive components (perception, memory, reasoning, decision-making).

Key capabilities required for AGI


A working list of abilities generally considered necessary (though definitions vary):

  1. Broad generalization and transfer:
  • Apply knowledge from one domain to new, different domains with minimal additional training.
  1. Efficient learning and meta-learning:
  • Rapidly learn new tasks from few examples or weak supervision.
  1. Abstraction and symbolic reasoning:
  • Form and manipulate high-level abstractions, causal models, and logical structures.
  1. Robust planning and long-horizon reasoning:
  • Plan over long temporal horizons under uncertainty and partial observability.
  1. Continual learning and memory:
  • Accumulate and reuse knowledge without catastrophic forgetting; maintain episodic and semantic memory.
  1. Multi-modal perception and integration:
  • Combine visual, auditory, textual, proprioceptive, and other signals coherently.
  1. Social, language, and commonsense understanding:
  • Communicate naturally, understand social cues, and reason about human goals and norms.
  1. Creativity and problem invention:
  • Generate novel solutions, hypotheses, and artifacts in new domains.
  1. Physical interaction and manipulation (for embodied AGI):
  • Robust motor skills, perception-action loops, and physical problem-solving.
  1. Self-monitoring, reflection, and learning about objectives:
  • Monitor internal states, reason about own performance, and adapt goals.

Architectural approaches and research paths


  1. Scaling up deep learning / foundation models
  • Approach: Train ever-larger models on massive, diverse datasets using self-supervised objectives (e.g., masked language modeling).
  • Rationale: Empirical success suggests scale yields emergent capabilities—broader knowledge, in-context learning, zero/few-shot performance.
  • Examples: GPT series, PaLM, LLaMA, and multi-modal variants.
  • Limitations: Data and compute costs, brittleness, lack of grounded causality, challenges with planning, reasoning, and sample efficiency.
  1. Neurosymbolic and hybrid systems
  • Approach: Combine neural networks for perception and pattern recognition with symbolic systems for reasoning and manipulation of explicit structures.
  • Rationale: Symbolic representations support systematicity, interpretability, and logical reasoning.
  • Research: Integrating logic programming, knowledge graphs, program synthesis with neural modules.
  1. Modular and cognitive-architecture approaches
  • Approach: Build architectures composed of interactive modules (perception, memory, planner, controller), possibly inspired by cognitive science.
  • Examples: SOAR, ACT-R, NARS; more modern modular deep architectures (e.g., modular RL).
  • Rationale: Modularity supports reusability, interpretability, and compositionality.
  1. Meta-learning and few-shot learning
  • Approach: Train systems to rapidly adapt to new tasks by learning update rules, priors, or initialization states (e.g., MAML).
  • Rationale: Enables efficient transfer and reduces data requirements for new tasks.
  1. Continual and lifelong learning
  • Methods: Elastic Weight Consolidation, memory replay, progressive networks, dynamically expandable architectures.
  • Goal: Maintain long-term competence across tasks and lifelong knowledge accumulation.
  1. Embodied and developmental approaches
  • Approach: Use robots or simulated agents to learn through interaction, exploration, and developmental curricula.
  • Rationale: Physical interaction grounds symbols, encourages causal model formation, and develops sensorimotor competence.
  1. Planning, search, and world models
  • Model-based RL, planning with learned models, Monte Carlo tree search, and program induction aim to combine learning with explicit forward models for planning.
  1. Causal and structured representation learning
  • Seek representations that reflect causal structure, enabling robust generalization under distribution shifts and interventions.

Benchmarks and evaluation


Evaluating AGI is challenging because AGI is not a single task; it must be assessed across breadth, adaptability, learning efficiency, and safety. Some evaluation approaches:

  • Behavioral / task-based tests:
  • Turing Test: Behavioral indistinguishability in conversation; criticized as narrow and anthropomorphic.
  • Abstraction and Reasoning Corpus (ARC): Tests generalization and abstract reasoning from few examples.
  • Animal-AI Testbed: Evaluates adaptive reasoning for embodied agents.
  • Big-bench and BIG-bench Hard: Wide arrays of language tasks probing reasoning and problem-solving.
  • GLUE / SuperGLUE: Benchmarks for natural language understanding (narrow relative to AGI).
  • MT-Bench / Model-based evaluation suites: Human-preference and multi-turn evaluation.
  • Formal / theoretical measures:
  • Universal intelligence (Legg & Hutter) gives a principled but impractical metric.
  • Performance-weighted evaluations across many environments with varying complexity and prior probabilities.
  • Practical evaluation axes:
  • Breadth: Number of distinct task families handled.
  • Efficiency: Data and compute required for new tasks.
  • Robustness: Performance under distribution shift, adversarial inputs, or degraded sensors.
  • Transfer and adaptation: Speed and quality of transfer to new tasks.
  • Safety and alignment: Propensity to follow instructions, avoid harmful behaviors, and respect constraints.

Current state of the art ...

Ready to see the full tree?

Clone the preview to open the complete learning structure, practice tools, and generated study materials.