System 2 Thinking: Definition and Examples
System 2 Thinking refers to a slow, deliberate, and analytical mode of reasoning, as opposed to System 1 (fast and intuitive). In prompt engineering, this concept is used to encourage AI models to produce more thoughtful responses by breaking down their reasoning step by step.
Full definition
System 2 Thinking is a concept from the work of psychologist and Nobel laureate Daniel Kahneman, presented in his book "Thinking, Fast and Slow" (2011). It distinguishes two cognitive systems: System 1, fast, automatic and intuitive, and System 2, slow, conscious and analytical. System 2 is mobilized when a task requires concentration, calculation or logical reasoning — for example solving a complex math problem or evaluating the validity of an argument.
In artificial intelligence and prompt engineering, the concept of System 2 Thinking has been adopted to describe techniques that push a language model to "think before responding". By default, an LLM generates text token by token sequentially, which resembles System 1: a fluid but sometimes superficial production. By introducing explicit step-by-step reasoning instructions (chain-of-thought), we force the model to simulate a deliberative process closer to System 2.
Techniques associated with System 2 Thinking include chain-of-thought prompting, self-consistency (generating multiple reasoning paths and selecting the most consistent), tree-of-thought (exploring multiple reasoning branches), and more recently "reasoning" models like o1 or Claude with extended thinking. These approaches significantly improve performance on logic, math, programming and complex analysis tasks.
The practical challenge is knowing when to activate this deep thinking mode. Not all queries require elaborate reasoning — a simple factual question does not need chain-of-thought. The prompt engineer must calibrate the level of thinking requested from the model based on task complexity, finding the right balance between response quality and computational cost.
Etymology
The term comes from the dual process theory of cognition, formalized by psychologists Daniel Kahneman and Amos Tversky in the 1970s-2000s. The names "System 1" and "System 2" were popularized by Kahneman in his best-seller "Thinking, Fast and Slow" (2011). The term was adopted in the AI field around 2022-2023 to describe deliberative reasoning approaches in LLMs.
Concrete examples
Solving a complex logic problem
Solve this problem step by step. Before giving your final answer, examine each hypothesis, verify your logic, and identify potential errors in your reasoning.
Critical analysis of an argument
Analyze the following argument using deliberate reasoning: first identify the premises, then evaluate their validity one by one, look for cognitive biases or fallacies, and finally formulate your conclusion.
Strategic decision-making
I need to choose between three options for my company. For each option, list the pros and cons, assess the risks, assign a score from 1 to 10 on each criterion, then recommend the optimal option justifying your choice.
Practical usage
To activate System 2 Thinking in your prompts, explicitly ask the model to break down its reasoning before concluding, using instructions like "think step by step" or "before answering, analyze each aspect of the problem". Reserve this approach for complex tasks (logic, math, strategic analysis) where deep reflection adds real value. For models that support it, enable native extended reasoning features rather than simulating the process via the prompt.
Related concepts
FAQ
What is the difference between System 1 and System 2 Thinking for an LLM?
Does System 2 Thinking always improve an AI's responses?
How do recent models like o1 or Claude natively integrate System 2?
See also
How to use this prompt
- Copy the prompt with the button above.
- Paste it into ChatGPT, Claude or your favorite AI assistant.
- Replace the bracketed variables with your details, then refine the result.
About Prompt Guide
Prompt Guide is a free library of 2500+ ready-to-use prompts for ChatGPT, Claude and other AIs, with guides to learn prompting and tools to build and optimize your own prompts.
More definitions
System Prompt: Definition and Examples
The system prompt is an initial hidden instruction, defined by the developer, that configures the behavior, tone, and limits of an AI model before
Take A Deep Breath: Definition and Examples
Prompt engineering technique that asks the language model to "take a deep breath" before answering, encouraging it to slow down and reason more carefully.
Temperature (AI): Definition and Examples
Temperature is a parameter that controls the degree of randomness and creativity in AI responses.
Test Time Compute: Definition and Examples
Test Time Compute refers to the computing power used by an AI model during inference (response generation), as opposed to the resources consumed during training.
Text Classification: Definition and Examples
Text classification is a natural language processing (NLP) technique that assigns one or more categories to a given text.
Text Summarization: Definition and Examples
Text summarization is an AI technique that condenses a long document into a shorter version while preserving the essential information and overall meaning.
Get new prompts every week
Join our newsletter.