thinking-hats-ai 1.0rc1

thinking-hats-ai is a Python package that facilitates idea generation by following Edward de Bono's Six Thinking Hats methodology from his Book. It enables you to generate ideas by selecting one of the six hats and lets you choose one of the implemented prompting technique to follow while generating the idea.

thinking-hats-ai: Python package implementing six thinking hats prompting

Package

What is it?

thinking-hats-ai is a Python package that facilitates idea generation by following Edward de Bono's Six Thinking Hats methodology from his Book. It enables you to generate ideas by selecting one of the six hats and lets you choose one of the implemented prompting technique to follow while generating the idea.

Table of Contents

• Use of Package
  • Example script
  • Hats
  • Prompting techniques
  • Brainstorming-input
  • Developer mode
• Installation
• Dependencies
• Background
• Creators

Use of Package

Example script

This example uses the CHAIN_OF_THOUGHT prompting techniques and the BLACK hat for the personality. It also uses developer mode to log the interaction in a separate file.

### Import package
from thinking_hats_ai import BrainstormingSession, Technique, Hat, BrainstormingInput

### Create session
session = BrainstormingSession('YOUR-OPENAI-API-KEY')
session.dev = True # Activate dev mode

### Define current status
brainstormingInput = BrainstormingInput(
    question = 'How could you make students come to class more often even though there are podcasts provided for each lecture?',
    ideas=[
        "Implement an interactive class participation system with incentives",
        "Extra credits or digital badges, encouraging students to attend and engage actively",
        "Offer exclusive in-class activities or discussions that are not available in the podcasts",
        "Create a social media group for students to share their experiences and insights from attending class",
        "Organize regular contests or challenges related to class content, with prizes for participants",
        "Provide a comfortable and engaging classroom environment with refreshments and seating arrangements",
        "Incorporate gamification elements into the class structure, such as quizzes or team-based activities",
    ],
    response_length='5 bullet points'
)

### Generate output
response = session.generate_idea(
    Technique.CHAIN_OF_THOUGHT,
    Hat.BLACK,
    brainstormingInput
)

### Print output
print(response)

Hats

The different hats act as a predefined persona according to Edward de Bono's book about the six thinking hats in brainstorming. You can select which persona should be used for your instance.

Hat	Role
WHITE	The White Hat represents neutrality and objectivity, focusing on gathering facts, identifying information gaps, and evaluating existing knowledge to ensure all reasoning is grounded in evidence and logic.
GREEN	The Green Hat represents creativity and innovation, focusing on generating new ideas, exploring alternatives, and proposing improvements to existing concepts to encourage original and unconventional thinking.
BLACK	The Black Hat represents critical judgment and caution, focusing on identifying potential risks, weaknesses, and negative outcomes of ideas to ensure they are practical and safe.
YELLOW	The Yellow Hat represents optimism and positivity, focusing on identifying the benefits, strengths, and potential value of ideas to highlight why they are worth pursuing.
BLUE	The Blue Hat represents control, organization, and overview, guiding the thinking process, setting objectives, managing the discussion, and ensuring that the Six Thinking Hats method is followed effectively.
RED	The Red Hat represents emotions, intuition, and gut feelings, allowing participants to express their instincts or emotional reactions to ideas without the need for justification or logic.

source: Book

Prompting techniques

The different prompting techniques help to analyse different approaches of idea generation for each hat. While implementing, we analyzed which of the techniques work best for which hat.

Technique	Explanation
CHAIN_OF_THOUGHT	The Chain of Thought technique guides the model to reason step-by-step, leveraging its advanced reasoning capabilities to improve coherence and depth in idea generation.
CHAIN_OF_VERIFICATION	The Chain of Verification technique refines ideas by chaining prompts, generating verification questions, and analyzing responses.
CHAINING	The Chaining technique builds a structured process: first understanding the thinking hat, then applying its perspective to the brainstorming context, and finally refining the response to ensure alignment.
CONTRASTIVE_PROMPTING	The Contrastive Prompting technique uses examples of good and bad responses to guide the agent towards an optimal output.
EMOTION_PROMPT	The EmotionPrompt technique enhances model performance by embedding emotional stimuli in prompts, encouraging more engaged, human-like, and context-aware responses.
FEW_SHOT	The Few Shot technique uses example-based prompting by first generating three example responses from the hat's perspective, then applying these patterns to produce a final, guided response.
MULTI_AGENT	The Multi-Agent technique involves multiple agents collaborating on a topic before converging on a refined final output.
PERSONA_PATTERN	The Persona Pattern technique generates ideas by adopting predefined persona, offering unique perspective and goals.
REACT	The ReAct (Reason and Act) technique combines reasoning and action by allowing models to interleave thought processes with tool use, enabling more accurate and interactive problem-solving.
SYSTEM_2_ATTENTION	The System 2 Attention technique introduces a two-step prompting process: first filtering and organizing brainstorming input, then using this optimized context to generate higher-quality ideas.
TAKE_A_STEP_BACK	The Take-a-Step-Back technique adds a follow-up question prompting the model to reflect on its initial reasoning before giving a final answer.
ZERO_SHOT_COT	The Zero-shot Chain-of-Thought technique guides the model to reason step-by-step, breaking down complex problems into intermediate steps for improved accuracy and transparency.

Brainstorming-input

The instance of BrainstormingInput allows you to pass the brainstorming question, ideasand response_length to the generation of an idea.

Variable Name	Explanation
question	This variable takes a string, the question that was asked in the brainstorming session
ideas	This variable takes a list of strings where each string is a idea from the brainstorming session
response_length	This variable takes a string which will control the length of the answer. You can say "10 sentences" but also things like "similar to the other ideas". It should fit the sentence: Provide a response with a length of {response_length}.

Developer mode

The developer mode is used to log the in/outputs of the api calls. A log folder and log files will be created when executing a Script with activated developer mode. This was implemented for prompt engineering purposes and help to analyse the history of all API calls made.

It can be activated by setting the dev attribute to True (default False).

instance.dev = True

Installation

This package is available through the Python Package Index (PyPI).

pip install thinking-hats-ai

Dependencies

• LangChain - A framework for developing applications powered by language models
• AG2 - A multi-agent framework for building AI-driven systems
• Autogen - A framework for enabling autonomous AI agents to collaborate

Background

The implementation of thinking-hats-ai started at UZH as a part of three bachelors theses.

Creators

• Timon Derboven - [e-mail]
• Leon Braga - [e-mail]
• Marc Huber - [e-mail]

---

Go to Top