Robert Greene's AI Law: Make Yourself Hard to Replace

Robert Greene's useful AI point is not that everyone should panic. It is sharper than that: AI changes the tools, but it does not remove power, ego, politics, dependence, trust, or human judgment from work.

In Calum Johnson's interview, Greene keeps returning to one practical law: create dependence. In the AI era, that means becoming the person people rely on for context, judgment, relationships, taste, systems, and calm execution when everyone else is using the same tools to produce the same average output.

Source Note

This post is based on Calum Johnson's interview with Robert Greene, the supplied transcript, Robert Greene's official channels, Penguin Random House's page for The Law of the Sublime, and public research from Microsoft, Anthropic, and the World Economic Forum on AI and work.

The psychological and workplace-power sections are educational commentary, not clinical, legal, HR, or mental-health advice. If a workplace situation involves harassment, discrimination, abuse, retaliation, or severe distress, use qualified professional support instead of only internet advice.

Link Map

The AI Law: Create Dependence

Greene points to the law of creating dependence as the most relevant power principle for the AI era. If your work is easy to describe, easy to measure, and easy to hand to a cheaper person or a machine, you are exposed. If your work carries rare judgment, relationships, context, and business consequence, you become harder to replace.

That does not mean hoarding information or making yourself a bottleneck. In a healthy company, "create dependence" should mean creating trusted leverage:

• You understand the customer better than the dashboard does.
• You know which AI output is good, wrong, risky, or off-brand.
• You connect departments that normally do not speak clearly to each other.
• You turn vague goals into shipped systems.
• You create documented workflows that others can run, but still trust you to improve.

In other words, the best version of dependence is not secrecy. It is becoming the person whose taste, context, and judgment improve the whole machine.

Power Still Matters After Automation

One of Greene's sharpest points is that even if AI reduces a 100-person company to a 10-person company, there will still be politics. There will still be bosses, egos, credit, fear, influence, and status. AI does not delete human nature.

That matters because many AI career conversations are too technical. They talk about prompts, agents, and tools, but not power. In real teams, the valuable person is not always the person who can generate the most text or code. It is often the person who can:

• set the right goal,
• protect the team from bad work,
• translate between business and technical people,
• keep evidence of their contribution,
• manage up without becoming invisible,
• use AI to make others more effective.

AI makes execution cheaper. That makes goal selection, taste, and internal trust more valuable.

The Irreplaceable Stack

If I were turning Greene's advice into a JQ AI SYSTEMS career framework, I would build the irreplaceable stack like this:

The pattern is simple: let AI expand your capacity, but keep ownership of meaning, trust, and judgment.

Difficult People, Credit, and Emotional Distance

A large part of the episode is about narcissistic bosses and manipulative people. Greene's practical phrase is "it's not personal." That does not excuse bad behavior. It creates distance. If someone is trying to pull you into drama, your first job is not to win the drama. It is to stop donating your attention to it.

For AI-era workers, this matters because speed can make credit and ownership messier. If everyone is generating decks, code, reports, and strategies with AI, the person who documents their work becomes harder to erase.

Practical moves:

• Keep a decision log for important work.
• Send short written recaps after meetings.
• Attach your name to the problem definition, not just the output.
• Store before-and-after evidence when you improve a workflow.
• Document the judgment calls AI could not make.
• Escalate serious workplace issues through proper HR, legal, or professional channels.

In a world where output is abundant, proof of judgment becomes a career asset.

If AI Targets Your Role

Greene's advice is blunt: first assess whether the tsunami is coming for your exact role or only part of your role. Those are different situations.

If AI is automating the core of your job, you may need to recreate yourself around adjacent skills. If AI is only making your field more competitive, your job is to identify what humans still do better and move toward that edge.

Good questions:

• Which parts of my work are repetitive, text-heavy, or rules-based?
• Which parts require trust, taste, negotiation, empathy, or accountability?
• Which customers still want a human because automation feels frustrating?
• Which workflows can I redesign with agents instead of waiting to be redesigned by management?
• What adjacent role uses my experience but adds more judgment?

Microsoft's Work Trend Index frames this as the rise of human-agent work. Anthropic's Economic Index shows AI already appearing across many work tasks. The practical response is not denial. It is role redesign.

Prompt Pack

Use these prompts to turn Greene's advice into a personal work audit.

1. Irreplaceability Audit

Act as my career strategist in the AI era.

My role:
[describe your role]

My weekly tasks:
[paste a short list]

Analyze:
1. Which tasks AI can already automate or compress
2. Which tasks still need human judgment, trust, context, or accountability
3. Where I am currently replaceable
4. Where I could create healthy dependence
5. What skill stack would make me harder to replace in 90 days

Be direct. Do not flatter me.

2. Create Dependence Without Becoming a Bottleneck

I want to become more valuable without hoarding knowledge or blocking the team.

Given this workflow:
[describe workflow]

Design a system where:
- AI handles repeatable execution
- I own judgment, review, context, and exceptions
- The team can see my contribution clearly
- The workflow is documented
- My value increases because I improve the system, not because I hide it

Return a concrete 4-week implementation plan.

3. Credit and Proof Log

Help me create a simple proof log for my work.

For each project, I want to track:
- Problem I identified
- Decision I made
- AI/tools I used
- Before state
- After state
- Business impact
- People involved
- Evidence links
- Next improvement

Create a lightweight template I can update weekly.

4. Recreate Yourself Around AI

My current role may be disrupted by AI.

Current role:
[role]

Skills I have:
[skills]

Work I enjoy:
[interests]

Constraints:
[location, income needs, time, education, family, etc.]

Suggest three adjacent career paths where my existing experience still matters, but where AI becomes leverage instead of replacement.
For each path, give me:
- Why it fits
- What to learn first
- One portfolio project
- One service I could offer
- First 30-day plan

Builder Checklist

• Stop measuring your value by volume of output.
• Map the tasks in your work that AI can already compress.
• Move toward judgment, taste, trust, and exception handling.
• Build agent workflows that make you more useful to the team.
• Document decisions so your contribution is visible.
• Keep learning tools, but do not become a tool collector.
• When work feels political, protect your attention and evidence trail.
• If your field is being heavily automated, recreate yourself early.

Sources

• The Calum Johnson Show: Robert Greene interview
• Calum Johnson on X
• Calum Johnson on Instagram
• The Calum Johnson Show on Instagram
• The Calum Johnson Show on YouTube
• Robert Greene Official on YouTube
• Robert Greene on Instagram
• Robert Greene on X
• Robert Greene official website
• Penguin Random House: The Law of the Sublime
• Microsoft Work Trend Index 2026
• Microsoft Work Trend Index 2025
• Anthropic Economic Index
• Anthropic: Labor market impacts of AI
• World Economic Forum: Future of Jobs Report 2025
• Zapier MCP