AI Displacement Monitor
Use this skill to produce a structured risk monitor for AI-led labor substitution and downstream financial stress.
Output Format
Always return:
- 1. Signal Board (10 indicators with latest value, direction, threshold status)
- Composite Risk Light (
GREEN / YELLOW / ORANGE / RED) - Actionable Notes (portfolio/risk posture suggestions)
- Data Gaps (missing or stale inputs)
Indicator Framework
Read references/thresholds.example.json and follow its indicator IDs, thresholds, and tiering.
Also apply the "Industrial-Revolution Lens" when interpreting risk:
- - Do not evaluate layoffs alone.
- Compare substitution speed vs re-absorption speed (new demand + new capex).
- If substitution weakens labor but capex/reinvestment accelerates, avoid over-escalating crisis labels.
- - Tier A (Leading labor demand): A1-A4
- Tier B (Labor market confirmation): B1-B3
- Tier C (Spillover: consumption/credit): C1-C3
Composite Rule
- - YELLOW: Tier A triggered >= 2
- ORANGE: Tier A >= 2 and Tier B >= 1
- RED: Tier A >= 2 and Tier B >= 1 and Tier C >= 1
- GREEN: otherwise
Weak-Links Interpretation (Jones Lens)
When assessing macro impact, apply a weak-links check:
- - Broad automation can still deliver gradual macro gains if key bottleneck tasks remain scarce.
- Do not infer immediate macro collapse from partial task automation alone.
- If bottleneck proxies remain tight (D3 worsening, D4 weak reinvestment), keep risk elevated.
- If bottlenecks ease via reinvestment/capex and purchasing power improves (D1/D2), avoid over-escalation.
Minimum Quality Rules
- - Time-stamp each metric and note frequency mismatch (weekly vs monthly vs quarterly).
- If source coverage is partial, mark confidence as
low or medium. - Never hide missing data; list it under Data Gaps.
- If more than 3 indicators are missing, downgrade confidence by one level.
Recommended Alert Style
Keep alerts short and decision-oriented:
- - "What changed"
- "Why it matters now"
- "What to do next"
Optional JSON Mode
If user asks for machine-readable output, return:
- - INLINECODE7
- INLINECODE8 (id, value, unit, threshold, triggered, trend)
- INLINECODE9
- INLINECODE10
- INLINECODE11
- INLINECODE12
AI 替代监控器
使用此技能生成AI主导的劳动力替代及下游金融压力的结构化风险监控器。
输出格式
始终返回:
- 1. 信号板(10项指标,含最新值、方向、阈值状态)
- 综合风险指示灯(绿色 / 黄色 / 橙色 / 红色)
- 可操作说明(投资组合/风险态势建议)
- 数据缺口(缺失或过时的输入项)
指标框架
读取 references/thresholds.example.json 并遵循其指标ID、阈值和分级标准。
在解读风险时还需应用工业革命视角:
- - 不要仅评估裁员情况。
- 比较替代速度与再吸收速度(新需求+新资本支出)。
- 若替代削弱劳动力但资本支出/再投资加速,避免过度升级危机标签。
- - A级(领先劳动力需求):A1-A4
- B级(劳动力市场确认):B1-B3
- C级(溢出效应:消费/信贷):C1-C3
综合规则
- - 黄色:A级触发 >= 2项
- 橙色:A级 >= 2项 且 B级 >= 1项
- 红色:A级 >= 2项 且 B级 >= 1项 且 C级 >= 1项
- 绿色:其他情况
薄弱环节解读(琼斯视角)
评估宏观影响时,应用薄弱环节检查:
- - 若关键瓶颈任务仍然稀缺,广泛自动化仍可带来渐进式宏观收益。
- 不要仅凭部分任务自动化推断立即的宏观崩溃。
- 若瓶颈代理指标保持紧张(D3恶化,D4再投资疲弱),维持风险高位。
- 若通过再投资/资本支出缓解瓶颈且购买力改善(D1/D2),避免过度升级。
最低质量标准
- - 为每项指标标注时间戳,并注明频率差异(周度 vs 月度 vs 季度)。
- 若数据源覆盖不完整,将置信度标记为低或中。
- 绝不隐藏缺失数据;将其列在数据缺口下。
- 若缺失超过3项指标,将置信度下调一级。
推荐警报风格
保持警报简短且以决策为导向:
可选JSON模式
若用户要求机器可读输出,返回:
- - asOf
- signals[](id, value, unit, threshold, triggered, trend)
- composite
- confidence
- gaps[]
- notes[]