AI Debate & Consensus

AI ↔ AI

Multiple AI analysts argue positions, system aggregates consensus.

6 nodes · 7 edgesresearch

agentsystem

Visual

Pose Research Questionsystem

Frame the question with context and evaluation criteria.

↓parallel→ Analyst A (Optimistic)

↓parallel→ Analyst B (Skeptical)

↓parallel→ Analyst C (Neutral)

Analyst A (Optimistic)agent

Argue for the opportunity, highlight upside potential.

↓parallel→ Aggregate Positions

Analyst B (Skeptical)agent

Argue against, surface risks and failure modes.

↓parallel→ Aggregate Positions

Analyst C (Neutral)agent

Weigh both sides objectively with evidence.

↓parallel→ Aggregate Positions

Aggregate Positionssystem

Score arguments by evidence quality and logical coherence.

↓sequential→ Generate Consensus Report

Generate Consensus Reportagent

Produce final report with majority view and dissenting opinions.

Open in Editor

uc-ai-debate.osop.yaml

osop_version: "1.0"
id: "ai-debate"
name: "AI Debate & Consensus"
description: "Multiple AI analysts argue positions, system aggregates consensus."

nodes:
  - id: "pose_question"
    type: "system"
    name: "Pose Research Question"
    description: "Frame the question with context and evaluation criteria."

  - id: "analyst_1"
    type: "agent"
    subtype: "llm"
    name: "Analyst A (Optimistic)"
    description: "Argue for the opportunity, highlight upside potential."

  - id: "analyst_2"
    type: "agent"
    subtype: "llm"
    name: "Analyst B (Skeptical)"
    description: "Argue against, surface risks and failure modes."

  - id: "analyst_3"
    type: "agent"
    subtype: "llm"
    name: "Analyst C (Neutral)"
    description: "Weigh both sides objectively with evidence."

  - id: "vote"
    type: "system"
    name: "Aggregate Positions"
    description: "Score arguments by evidence quality and logical coherence."

  - id: "report"
    type: "agent"
    subtype: "llm"
    name: "Generate Consensus Report"
    description: "Produce final report with majority view and dissenting opinions."

edges:
  - from: "pose_question"
    to: "analyst_1"
    mode: "parallel"
  - from: "pose_question"
    to: "analyst_2"
    mode: "parallel"
  - from: "pose_question"
    to: "analyst_3"
    mode: "parallel"
  - from: "analyst_1"
    to: "vote"
    mode: "parallel"
  - from: "analyst_2"
    to: "vote"
    mode: "parallel"
  - from: "analyst_3"
    to: "vote"
    mode: "parallel"
  - from: "vote"
    to: "report"
    mode: "sequential"

←

Model Cascade / Fallback Chain

Autonomous Code Fix Pipeline

→