🧠Artificial General Intelligence

NEW

Also known as: AGI

Hypothetical AI that achieves broad human-level competence across diverse tasks, generalizes, and adapts without task-specific engineering.

DESCRIPTION

Artificial General Intelligence refers to systems capable of human-level or superior performance across most tasks, including reasoning, learning, creativity, and transfer. Unlike narrow AI optimized for specific problems, AGI would generalize widely and adapt to novel domains. Discussions focus on feasibility, timelines, and safety, including alignment and control. AGI remains hypothetical, with research exploring cognitive architectures, meta-learning, and scaling laws. Governance and ethics are central to mitigate potential societal risks.

ALTERNATIVE DEFINITIONS

Wikipedia

Artificial general intelligence (AGI) is the hypothetical ability of an intelligent agent to understand or learn any intellectual task that a human being can.

Defines AGI by universal task capability matching humans.

Source

OpenAI

Artificial general intelligence (AGI) is highly autonomous systems that outperform humans at most economically valuable work.

Emphasizes autonomy and economic performance as benchmarks.

Source

RELATED TERMS

Artificial Intelligence Artificial Super Intelligence Reinforcement Learning Cognitive Architecture Superclass Autonomous Agent AI Safety Alignment

COMPETING TERMS

Artificial Narrow Intelligence

RELATIONS

• Broader: Artificial Intelligence

KNOWLEDGE AREAS

AI AI Governance Data Science

TARGET AUDIENCE

AI Engineer Data Scientist Chief Technology Officer Chief Data Officer AI Product Manager

EXTERNAL SOURCES

• Wikipedia - Artificial general intelligence
• Stanford HAI - Overview and implications of AGI
• Alignment Forum - AI alignment discussions

TOP QUESTIONS

Why is Artificial General Intelligence important?

AGI could transform economies and sciences, solving complex problems beyond narrow AI. Its potential impact heightens focus on safety, governance, and societal alignment.

Category: Relevancy

Is Artificial General Intelligence achievable with current techniques?

Opinions vary; some believe scaling current models will reach AGI, others argue new breakthroughs in reasoning and memory are needed.

Category: Opinion

What risks are associated with Artificial General Intelligence?

AGI poses risks of misalignment, unintended consequences, concentration of power, and automation shocks. Robust governance and technical safety are essential.

Category: Risk

How does Artificial General Intelligence differ from Artificial Narrow Intelligence?

ANI excels in specific tasks; AGI generalizes across domains and adapts to novelty without substantial task-specific engineering.

Category: Difference

What learning approaches are relevant for Artificial General Intelligence?

Meta-learning, reinforcement learning, planning, memory architectures, and multimodal representation learning are often discussed for AGI capabilities.

Category: Approach

What governance frameworks matter for Artificial General Intelligence?

AI safety research, risk management frameworks (e.g., NIST AI RMF), model cards, and international collaboration guide responsible development.

Category: Governance

Who coined the term Artificial General Intelligence?

The term gained popularity in the 2000s; while not attributed to a single inventor, it was used by researchers including Ben Goertzel.

Category: Origin

What is a practical milestone toward Artificial General Intelligence?

Consistent cross-domain reasoning, reliable tool use, and long-horizon planning under uncertainty would be significant markers.

Category: Milestone

How should organizations prepare for Artificial General Intelligence?

Develop responsible AI policies, invest in safety research, upskill teams, and participate in multi-stakeholder governance efforts.

Category: Preparation