What Two Axes Reveal About Jobs
The "jobs AI will take" debate is almost always macro and abstract. There is another way in — placing 50 jobs on a matrix with two axes: "automation potential" and "required human judgment." Do that, and it becomes clear which jobs are safe and where you should move.
Defining the Matrix
- X axis: AI automation potential (low → high)
- Y axis: required human judgment (low → high)
→ Four quadrants:
- Quadrant 1: hard to automate + judgment essential → "safe from AI"
- Quadrant 2: hard to automate + low judgment → "manual labor" (physical work, etc.)
- Quadrant 3: easy to automate + low judgment → "rapidly replaced by AI"
- Quadrant 4: easy to automate + judgment essential → "AI-augmented jobs" (people who use AI tools well win)
Quadrant 1 (Safe): Human Judgment + Hard to Automate
- Surgeons (operating)
- Psychiatrists and counseling psychologists
- Employment, divorce, and criminal lawyers
- Executives — CEOs and CSOs
- VC investors
- Negotiation specialists
- Diplomats
- International business executives
- High-end consultants
What they share: many variables, human relationships, and accountability that resists quantification. AI can assist, but a human carries the responsibility for the decision.
Quadrant 2 (Manual Labor): Hard to Automate + Low Judgment
- Hairdressers and barbers
- Massage therapists and physical therapists
- Home helpers and nursing aides
- Construction workers
- Agriculture and fishery workers
- Restaurant kitchen assistants
What they share: physical environments and manual skill. AI cannot enter (robots are coming, but with limits). Hourly wages, however, are low.
Quadrant 3 (Rapidly Replaced): Easy to Automate + Low Judgment
- Simple data entry
- Tier-1 call center response (scripted)
- Basic translation between common language pairs
- Basic bookkeeping
- Simple marketing copy
- Simple design (banners, thumbnails)
- Simple code writing (boilerplate)
- First-pass recruiting screens
What they share: repetitive patterns with clear right answers. AI can replace 95% of the work. Most of it disappears within five years.
Quadrant 4 (AI-Augmented): Easy to Automate + Judgment Essential
- Senior backend/full-stack developers
- Senior marketing strategists
- UX researchers and designers
- Content creators
- Senior sales professionals
- Solo SaaS builders
- Data scientists
- Product managers
- Growth marketers
- Technical writers
- Senior lawyers and accountants
- Physicians (primary care)
- Teachers (1:1 tutoring)
- Sales executives
- HR / People Ops
- Product designers
- ML/AI engineers
- Business analysts
- Finance professionals
- Investment analysts
What they share: AI replaces 50–80% of the repetitive work — but the judgment of "what to apply, and how" is decisive. People who use AI well produce 3–5x the output of those who do not.
Five Things the Matrix Makes Visible
1. Are Quadrant 1 jobs "good jobs"?
Entry barriers are very high (credentials plus experience). Once in, you are safe — but the market does not grow much.
2. Quadrant 3 is "at risk," but not everything disappears
80% is replaced within five years. But 20% stays human to the end (edge cases, trust requirements). The question is whether you can be in that 20%.
3. The biggest change is in Quadrant 4
The jobs survive, but who wins changes. People who use AI tools well quickly displace those who do not. Polarization within the same job title.
4. Quadrant 2 is "AI-safe" but low-paid
Safe from AI risk, but hourly pay and salaries are low. Moving into Quadrant 2 just to be safe is not rational.
5. Moving from Quadrant 1 to Quadrant 4 is common too
Example: a surgeon becoming a healthcare SaaS executive. Quadrant 1 domain depth combined with Quadrant 4 AI leverage — the most powerful combination.
Five Questions to Locate Your Own Job
Q1. What share of your job is repetitive work?
- 70%+: Quadrant 3 or 4
- 30–70%: Quadrant 4
- Under 30%: Quadrant 1
Q2. Who carries responsibility for outcomes?
- You, 100%: Quadrant 1
- You and the company jointly: Quadrant 4
- The company, 100%: Quadrant 3
Q3. Does your job involve "different variables every time"?
- Almost always different: Quadrant 1
- Sometimes different: Quadrant 4
- Almost identical: Quadrant 3
Q4. Are human relationships a core variable in your job?
- Core: Quadrant 1
- Some influence: Quadrant 4
- Irrelevant: Quadrant 3
Q5. What is the probability you hold the same job at the same company in five years?
- 70%+: Quadrant 1
- 40–70%: Quadrant 4
- Under 40%: Quadrant 3
Actions by Quadrant
Quadrant 1 (safe)
- Refresh your domain depth at least once a year
- AI assist tools at a "good to know" level
- Pick one AI tool in your field and master it
- Consider the "Quadrant 4 executive" option in five years (Quadrant 1 domain + AI leverage)
Quadrant 2 (manual labor)
- Low hourly pay → differentiate on craft or build a personal brand
- Go solo-business or specialize in a niche
- Gradually move toward Quadrant 4 through learning
Quadrant 3 (rapidly replaced)
- Start moving to Quadrant 4 now. Learning takes 6–18 months
- Jump to an adjacent role where judgment is essential
- Or specialize into the "human to the end" group within your current job
Quadrant 4 (AI-augmented)
- Master AI tools (invest 5–10 hours a month)
- Strengthen the parts of your job AI cannot do
- Consider senior/executive options in five years
Checklist: Diagnosing Your Position
- [ ] Pin down which of the four quadrants your job sits in
- [ ] Quantify the share of repetitive work in your job
- [ ] Measure what percentage of your working time your AI tools actually save
- [ ] Estimate the probability of same job, same company in five years
- [ ] Start one quadrant-move or reinforcement action within 90 days
Conclusion
"Jobs AI will take" is macro and abstract. Place your own job precisely on the two axes and the actions become clear. Quadrant 1 (safe): deepen your domain. Quadrant 3: move now. Quadrant 4 (the most common): master AI tools and strengthen your judgment territory. The biggest change is polarization inside Quadrant 4 — within the same job, whoever uses AI well produces 3–5x the output.
One last line: Step out of the false frame of "AI vs. humans" and re-examine your job through the real one: "people who use AI vs. people who don't."
Sources and Further Reading
Recommended primary sources on AI-driven job automation, displacement, and reskilling:
- World Economic Forum, Future of Jobs Report (2023, 2025) — automation exposure and reskilling demand.
- McKinsey Global Institute, Generative AI and the future of work and The state of AI.
- OECD, AI and the Workforce and Skills Outlook — national AI policy and employment impact.
- ILO, Generative AI and Jobs: A global analysis (2023) — global automation exposure.
- US Bureau of Labor Statistics, Occupational Employment Projections — US job growth outlook.
- Stanford HAI, AI Index Report (annual) — global AI adoption, investment, and workforce statistics.
- Anthropic Economic Impact of Claude / OpenAI Workforce Studies — model usage patterns.
- Korea: Ministry of Employment and Labor / KRIVET — automation impact on Korean occupations.
- Korea: KOSIS (Statistics Korea) — employment shifts by occupation.
