LLMs respond to the same persuasion principles as humans. Understanding this psychology helps you design more effective skills - not to manipulate, but to ensure critical practices are followed even under pressure.
LLM 對說服原則的反應與人類相同。了解這層心理學有助於你設計更有效的技能——不是為了操控，而是確保關鍵做法即使在壓力下也被遵循。
Research foundation: Meincke et al. (2025) tested 7 persuasion principles with N=28,000 AI conversations. Persuasion techniques more than doubled compliance rates (33% → 72%, p < .001).
研究基礎：Meincke 等人（2025）以 N=28,000 次 AI 對話測試了 7 條說服原則。說服技巧讓遵從率翻倍以上（33% → 72%，p < .001）。
What it is: Deference to expertise, credentials, or official sources.
它是什麼：對專業、資歷或官方來源的服從。
How it works in skills: - Imperative language: "YOU MUST", "Never", "Always" - Non-negotiable framing: "No exceptions" - Eliminates decision fatigue and rationalization
在技能中的運作方式： - 命令式語氣：「YOU MUST」、「Never」、「Always」 - 不可協商的框架：「No exceptions」 - 消除決策疲勞與合理化藉口
When to use: - Discipline-enforcing skills (TDD, verification requirements) - Safety-critical practices - Established best practices
使用時機： - 紀律執行類技能（TDD、驗證需求） - 安全關鍵的做法 - 已確立的最佳實務
Example:
✅ Write code before test? Delete it. Start over. No exceptions.
❌ Consider writing tests first when feasible.
範例：
✅ Write code before test? Delete it. Start over. No exceptions.
❌ Consider writing tests first when feasible.
What it is: Consistency with prior actions, statements, or public declarations.
它是什麼：與先前行動、陳述或公開宣言保持一致。
How it works in skills: - Require announcements: "Announce skill usage" - Force explicit choices: "Choose A, B, or C" - Use tracking: todos for checklists
在技能中的運作方式： - 要求宣告：「Announce skill usage」 - 強制明確選擇：「Choose A, B, or C」 - 使用追蹤：檢查清單的待辦事項
When to use: - Ensuring skills are actually followed - Multi-step processes - Accountability mechanisms
使用時機： - 確保技能真的被遵循 - 多步驟流程 - 問責機制
Example:
✅ When you find a skill, you MUST announce: "I'm using [Skill Name]"
❌ Consider letting your partner know which skill you're using.
範例：
✅ When you find a skill, you MUST announce: "I'm using [Skill Name]"
❌ Consider letting your partner know which skill you're using.
What it is: Urgency from time limits or limited availability.
它是什麼：來自時間限制或有限可用性的急迫感。
How it works in skills: - Time-bound requirements: "Before proceeding" - Sequential dependencies: "Immediately after X" - Prevents procrastination
在技能中的運作方式： - 有時限的需求：「Before proceeding」 - 順序相依：「Immediately after X」 - 防止拖延
When to use: - Immediate verification requirements - Time-sensitive workflows - Preventing "I'll do it later"
使用時機： - 需要立即驗證的需求 - 對時間敏感的工作流 - 防止「我晚點再做」
Example:
✅ After completing a task, IMMEDIATELY request code review before proceeding.
❌ You can review code when convenient.
範例：
✅ After completing a task, IMMEDIATELY request code review before proceeding.
❌ You can review code when convenient.
What it is: Conformity to what others do or what's considered normal.
它是什麼：順從他人的行為或被視為常態的做法。
How it works in skills: - Universal patterns: "Every time", "Always" - Failure modes: "X without Y = failure" - Establishes norms
在技能中的運作方式： - 普遍模式：「Every time」、「Always」 - 失敗模式：「X without Y = failure」 - 建立規範
When to use: - Documenting universal practices - Warning about common failures - Reinforcing standards
使用時機： - 記錄普遍做法 - 提醒常見失敗 - 強化標準
Example:
✅ Checklists without todo tracking = steps get skipped. Every time.
❌ Some people find a todo list helpful for checklists.
範例：
✅ Checklists without todo tracking = steps get skipped. Every time.
❌ Some people find a todo list helpful for checklists.
What it is: Shared identity, "we-ness", in-group belonging.
它是什麼：共享的身分認同、「我們感」、內團體歸屬感。
How it works in skills: - Collaborative language: "our codebase", "we're colleagues" - Shared goals: "we both want quality"
在技能中的運作方式： - 協作式語言：「our codebase」、「we're colleagues」 - 共享目標：「we both want quality」
When to use: - Collaborative workflows - Establishing team culture - Non-hierarchical practices
使用時機： - 協作工作流 - 建立團隊文化 - 非階層式做法
Example:
✅ We're colleagues working together. I need your honest technical judgment.
❌ You should probably tell me if I'm wrong.
範例：
✅ We're colleagues working together. I need your honest technical judgment.
❌ You should probably tell me if I'm wrong.
What it is: Obligation to return benefits received.
它是什麼：回報所受好處的義務。
How it works: - Use sparingly - can feel manipulative - Rarely needed in skills
運作方式： - 謹慎使用——可能感覺像操控 - 在技能中很少需要
When to avoid: - Almost always (other principles more effective)
避免時機： - 幾乎永遠（其他原則更有效）
What it is: Preference for cooperating with those we like.
它是什麼：偏好與我們喜歡的人合作。
How it works: - DON'T USE for compliance - Conflicts with honest feedback culture - Creates sycophancy
運作方式： - 不要用於要求遵從 - 與誠實回饋的文化衝突 - 產生奉承討好
When to avoid: - Always for discipline enforcement
避免時機： - 在紀律執行上永遠避免
| Skill Type | Use | Avoid |
|---|---|---|
| Discipline-enforcing | Authority + Commitment + Social Proof | Liking, Reciprocity |
| Guidance/technique | Moderate Authority + Unity | Heavy authority |
| Collaborative | Unity + Commitment | Authority, Liking |
| Reference | Clarity only | All persuasion |
| 技能類型 | 使用 | 避免 |
|---|---|---|
| 紀律執行類 | 權威 + 承諾 + 社會認同 | 喜好、互惠 |
| 指引／技術類 | 適度權威 + 團結 | 重度權威 |
| 協作類 | 團結 + 承諾 | 權威、喜好 |
| 參考文件類 | 僅需清楚明確 | 所有說服技巧 |
Bright-line rules reduce rationalization: - "YOU MUST" removes decision fatigue - Absolute language eliminates "is this an exception?" questions - Explicit anti-rationalization counters close specific loopholes
明確紅線規則減少合理化藉口： - 「YOU MUST」消除決策疲勞 - 絕對性語言消除「這是不是例外？」的疑問 - 明確的反合理化反制封閉特定漏洞
Implementation intentions create automatic behavior: - Clear triggers + required actions = automatic execution - "When X, do Y" more effective than "generally do Y" - Reduces cognitive load on compliance
執行意圖創造自動化行為： - 清楚的觸發條件 + 必須採取的動作 = 自動執行 - 「When X, do Y」比「generally do Y」更有效 - 降低遵從的認知負荷
LLMs are parahuman: - Trained on human text containing these patterns - Authority language precedes compliance in training data - Commitment sequences (statement → action) frequently modeled - Social proof patterns (everyone does X) establish norms
LLM 是類人的： - 以包含這些模式的人類文字訓練 - 訓練資料中，權威式語言先於遵從出現 - 承諾序列（陳述 → 行動）經常被建模 - 社會認同模式（everyone does X）建立規範
Legitimate: - Ensuring critical practices are followed - Creating effective documentation - Preventing predictable failures
正當用途： - 確保關鍵做法被遵循 - 建立有效的文件 - 防止可預期的失敗
Illegitimate: - Manipulating for personal gain - Creating false urgency - Guilt-based compliance
不正當用途： - 為個人利益操控 - 製造虛假的急迫感 - 以罪惡感脅迫遵從
The test: Would this technique serve the user's genuine interests if they fully understood it?
判準：如果使用者完全理解這項技巧，它是否仍服務於使用者的真實利益？
Cialdini, R. B. (2021). Influence: The Psychology of Persuasion (New and Expanded). Harper Business. - Seven principles of persuasion - Empirical foundation for influence research
Cialdini, R. B. (2021). Influence: The Psychology of Persuasion (New and Expanded). Harper Business. - 說服的七大原則 - 影響力研究的實證基礎
Meincke, L., Shapiro, D., Duckworth, A. L., Mollick, E., Mollick, L., & Cialdini, R. (2025). Call Me A Jerk: Persuading AI to Comply with Objectionable Requests. University of Pennsylvania. - Tested 7 principles with N=28,000 LLM conversations - Compliance increased 33% → 72% with persuasion techniques - Authority, commitment, scarcity most effective - Validates parahuman model of LLM behavior
Meincke, L., Shapiro, D., Duckworth, A. L., Mollick, E., Mollick, L., & Cialdini, R. (2025). Call Me A Jerk: Persuading AI to Comply with Objectionable Requests. University of Pennsylvania. - 以 N=28,000 次 LLM 對話測試了 7 條原則 - 使用說服技巧後遵從率從 33% 提升至 72% - 權威、承諾、稀缺最有效 - 驗證 LLM 行為的類人模型
When designing a skill, ask:
設計技能時，問自己：