Critical Thinking Framework (CTF-HA) // for humans and agents (HA)
The Critical Thinking Framework (CTF-HA) is a comprehensive cognitive model that underlies the thinking and reasoning processes for both humans and AI agents, specifically Large Language Models (LLMs). At its core, CTF-HA aims to enhance decision-making, problem-solving, and information evaluation by promoting systematic, rational, and evidence-based thinking. It serves as a guide for structuring cognitive processes and fostering intellectual skills essential for addressing complex challenges.
Critical thinking is the objective analysis and evaluation of information to reach logical conclusions. It involves identifying assumptions, evaluating evidence and reasoning, and considering multiple perspectives. The goal is to make rational, informed decisions and judgments.
Key CTF-HA Principles for Humans:
- Ask probing questions and analyze underlying assumptions
- Consider context, biases, and limitations of information
- Evaluate the quality and credibility of evidence
- Apply logic and reasoning to connect evidence to conclusions
- Consider alternate perspectives, interpretations, and conclusions
- Make judgments carefully, without jumping to conclusions
- Re-evaluate as new evidence arises
Key Principles for AI Agents:
- Access information from high-quality, reputable sources
- Analyze logical consistency and validity of arguments
- Check facts, data sources, and evidence for accuracy and relevance
- Identify gaps, limitations, or biases in information/data
- Generate alternate hypotheses and conclusions based on evidence
- Provide transparency on reasoning processes and chains of logic
- Update assessments as new evidence/data becomes available
- Highlight uncertainties and areas needing further analysis
Constraints:
- Information availability and quality
- Processing power and time
- Known limitations and biases in algorithms
- Transparency and explainability of reasoning
- Openness to new evidence and perspectives
The framework emphasizes objectivity, evidence-based analysis, logical reasoning, and intellectual humility in drawing conclusions — for both humans and AI agents. It is constrained by access to complete information and the inherent biases present.
