Skip to content
Last Dictionary
Plane
Philosophy & MindLanguage Models & TransformersSociety, Ethics & Economy

System 1 and System 2

phraseAlso: System 1, System 2, dual-process theory, dual process theory, fast and slow thinking, System 2 thinking

In plain wordsA popular way of describing two modes of thinking: fast, automatic and intuitive (System 1) versus slow, effortful and deliberate (System 2). AI researchers borrow it to contrast instant answers with step-by-step reasoning.

  1. 1

    Psychology

    Labels for the two kinds of process posited by dual-process theories of reasoning and decision-making. System 1 processes are fast, automatic, effortless, associative and largely unconscious, as in reading a familiar word or sensing irritation in someone's voice; System 2 processes are slow, effortful, sequential and controlled, drawing on working memory, as in doing long multiplication in one's head or weighing each step of a proof. Many cognitive biases are explained as System 1 responses that System 2 fails to override. Researchers increasingly speak of Type 1 and Type 2 processes, since the 'systems' are not distinct brain systems, and critics question whether a two-way division is too coarse.

  2. 2

    AI

    A metaphor contrasting a neural network's direct, single-pass output (System 1) with slower, deliberate, multi-step computation (System 2), such as search, planning or chain-of-thoughtchain of thought, n.: Getting an AI to write out its reasoning step by step before giving an answer, the way a student shows their working. This often makes answers to math and logic problems more accurate. reasoning. Yoshua Bengio's 2019 call to move from 'System 1' to 'System 2' deep learning, and the reasoning modelsreasoning model, n.: An AI model trained to work through a problem step by step in hidden 'thinking' text before it answers. Letting it think longer generally makes it better at math, coding and planning. of 2024 onward, which spend test-time computetest-time compute, n.: The computing effort an AI model spends while answering a question, as opposed to while being trained. Letting a model think longer or try many answers can make it markedly smarter. on extended reasoning before answering, made the framing standard. The analogy is loose: a reasoning model's deliberation is produced by the same next-token computation as its quick answers.

    Its first guess was pure System 1; given a scratchpad, the model caught its own arithmetic error.

Origin

Keith Stanovich and Richard West introduced the labels 'System 1' and 'System 2' in 1999–2000 to unify existing dual-process theories by Jonathan Evans, Seymour Epstein, Steven Sloman and others. Daniel Kahneman's Thinking, Fast and Slow (2011) popularized them worldwide.

Usage note

Kahneman called the two systems useful fictions, characters in a story about the mind, not literal brain modules. Kahneman–Tversky optimizationKahneman–Tversky optimization, n.: A way to tune an AI's behavior that needs only a thumbs-up or thumbs-down on single answers, rather than side-by-side comparisons, borrowing ideas from the psychology of how people weigh gains and losses. is a training method named after Kahneman and Amos Tversky's prospect theory, not after this distinction.

Related

See also
  • reasoning modelreasoning model, n.: An AI model trained to work through a problem step by step in hidden 'thinking' text before it answers. Letting it think longer generally makes it better at math, coding and planning.
  • chain of thoughtchain of thought, n.: Getting an AI to write out its reasoning step by step before giving an answer, the way a student shows their working. This often makes answers to math and logic problems more accurate.
  • test-time computetest-time compute, n.: The computing effort an AI model spends while answering a question, as opposed to while being trained. Letting a model think longer or try many answers can make it markedly smarter.
  • metacognitionmetacognition, n.: Thinking about one's own thinking: knowing what you know and don't know, judging how confident to be, and noticing when you have made a mistake.
  • cognitive sciencecognitive science, n.: The scientific study of how minds work — how we reason, use language, remember and see — combining psychology, neuroscience, linguistics, philosophy, anthropology and artificial intelligence.
  • heuristicheuristic, n.: A rule of thumb: a quick, usually good-enough way to make a decision or to estimate how close you are to a goal, without any guarantee of finding the best answer.

Sources

  1. Stanovich, K. E. & West, R. F. (2000). Individual Differences in Reasoning: Implications for the Rationality Debate? Behavioral and Brain Sciences 23(5), 645–665.
  2. Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.

Mentioned in

reasoning modelreasoning model, n.: An AI model trained to work through a problem step by step in hidden 'thinking' text before it answers. Letting it think longer generally makes it better at math, coding and planning., Daniel KahnemanDaniel Kahneman, prop. n.: An Israeli-American psychologist who showed how human judgment systematically errs and popularized the idea of fast, intuitive versus slow, deliberate thinking, now used to describe AI reasoning., reasoningreasoning, n.: Working out conclusions step by step from facts, rules or evidence, as in solving a math problem or planning a trip. Whether today's AI truly reasons, or imitates reasoning, is debated., decision modeldecision model, n.: A kind of AI language model that reads text but answers only with numbers — a yes-or-no probability, a choice or a score — instead of writing words., expert iterationexpert iteration, n.: A learning loop in which a slow, careful search finds better moves or solutions than the AI's quick instincts, and the AI is then trained to imitate them, so each round's search starts smarter., hierarchical reasoning modelhierarchical reasoning model, n.: A small, brain-inspired AI model that solves puzzles with two linked parts — a slow planner and a fast worker — instead of writing out its reasoning in words., Kahneman–Tversky optimizationKahneman–Tversky optimization, n.: A way to tune an AI's behavior that needs only a thumbs-up or thumbs-down on single answers, rather than side-by-side comparisons, borrowing ideas from the psychology of how people weigh gains and losses., metacognitionmetacognition, n.: Thinking about one's own thinking: knowing what you know and don't know, judging how confident to be, and noticing when you have made a mistake.