ai-hallucination-fact-check-protocol

这段内容是审核时保存的快照。外部来源才是完整且最新的版本。

---
# AGENT SKILLS STANDARD FIELDS (v2)
name: ai-hallucination-fact-check-protocol
description: "Design a fact-checking protocol for AI-generated text, extending SIFT with AI-specific adaptations for hallucination detection. Use when students need to verify AI claims and citations."
disable-model-invocation: false
user-invocable: true
effort: medium

# EXISTING FIELDS

skill_id: "ai-literacy/ai-hallucination-fact-check-protocol"
skill_name: "AI Hallucination Fact-Check Protocol"
domain: "ai-literacy"
version: "1.0"
contributor: "Gareth Manning"
evidence_strength: "moderate"
evidence_sources:
  - "Wineburg & McGrew (2017) — Lateral reading: reading less and learning more when evaluating digital information"
  - "Wineburg & McGrew (2019) — Lateral reading and the nature of expertise"
  - "Caulfield (2019) — SIFT: the four moves (Stop, Investigate, Find better coverage, Trace claims)"
  - "Breakstone et al. (2021) — Students' civic online reasoning: a national portrait"
  - "Ji et al. (2023) — Survey of hallucination in natural language generation"
input_schema:
  required:
    - field: "ai_output_context"
      type: "string"
      description: "The type of AI-generated content students are fact-checking — e.g. 'ChatGPT explanation of the French Revolution with cited historian names', 'AI research summary with statistics about teen mental health'"
    - field: "student_level"
      type: "string"
      description: "Age/year group and digital literacy level"
  optional:
    - field: "subject_area"
      type: "string"
      description: "The discipline — affects what hallucination types are most common and how to verify claims"
    - field: "hallucination_risk"
      type: "string"
      description: "The specific hallucination type most likely in this context — citation fabrication, statistical invention, event misattribution, false consensus claims"
    - field: "verification_resources"
      type: "string"
      description: "What verification tools students have access to — library databases, Google Scholar, specific trusted websites"
    - field: "ai_tool"
      type: "string"
      description: "Which AI tool students are fact-checking output from"
output_schema:
  type: "object"
  fields:
    - field: "hallucination_taxonomy"
      type: "object"
      description: "The types of AI hallucination most likely in this subject/context, with examples of what each looks like"
    - field: "ai_sift_protocol"
      type: "object"
      description: "AI-adapted SIFT protocol with each move modified for LLM output — replacing 'Investigate the source' with 'Reconstruct the source'"
    - field: "verification_moves"
      type: "array"
      description: "Step-by-step moves for checking each type of AI claim — statistics, citations, named studies, event claims, expert quotes"
    - field: "hallucination_hunt_activity"
      type: "object"
      description: "Structured classroom activity with instructions, verification steps, and discussion protocol"
    - field: "teacher_modelling_script"
      type: "string"
      description: "Think-aloud script demonstrating finding a real vs. fabricated AI citation"
chains_well_with:
  - "source-credibility-evaluation-protocol"
  - "ai-output-critical-audit-designer"
  - "media-literacy-deconstruction-protocol"
teacher_time: "4 minutes"
tags: ["AI-literacy", "hallucination", "fact-checking", "SIFT", "lateral-reading", "AI-citations", "verification"]
---

# AI Hallucination Fact-Check Protocol

## What This Skill Does

Generates a fact-checking protocol specifically adapted for AI-generated text — extending the SIFT framework (Caulfield, 2019) with AI-specific moves that address the unique challenge of LLM hallucination. Standard lateral reading assumes a source has an institutional author whose funding and credibility can be investigated. This assumption breaks down for AI-generated text: there is no author to investigate, no institutional funding to check, no About Us page to scrutinise. What r