The Numbers

0%
of global jobs are exposed to AI
IMF, 2024
0%
of jobs in advanced economies exposed
IMF, 2024
0
net new jobs projected by 2030
WEF, 2025
increase in generative AI job postings (US, 2023→2024)
Stanford AI Index, 2025
0%
of all jobs face disruption by 2030 (displaced or newly created)
WEF Future of Jobs, 2025
0%
of employers plan AI-driven workforce reductions by 2030
WEF Future of Jobs, 2025
Key insight: Exposure to AI does not mean replacement. The IMF found that roughly half of exposed jobs in advanced economies are expected to benefit through productivity gains — not disappear.

Which Jobs Are Most Affected?

High Exposure

  • Finance & Accounting
  • Legal Services
  • Customer Support
  • Data Analysis
  • Journalism
  • Translation

Medium Exposure

  • Healthcare (diagnosis support)
  • Education
  • Marketing
  • Software Development
  • Architecture

Low Exposure

  • Skilled Trades
  • Nursing & Caregiving
  • Psychotherapy
  • Creative Direction
  • Construction

Augmentation vs Automation

Augmentation

AI assists the worker. The human stays in the role but works faster and on higher-value tasks. The ILO estimates this is 6× more likely than full automation.

Automation

AI replaces the worker entirely on a task or role. Most likely in highly routine, data-heavy jobs with clear rules and no need for physical presence or social judgment.

The Expert Debate

Optimistic — WEF / McKinsey

Net +78 million jobs by 2030. Generative AI could add $2.6–4.4 trillion in annual economic value. Productivity gains benefit the most exposed workers most.

Cautious — Acemoglu (MIT)

Only about 20% of US tasks are actually automatable at a profit. AI's true GDP contribution may be as low as 0.5% over the next decade. Transition costs for displaced workers are real and fast.

Both views come from serious economists using the same underlying data. Forming your own view — carefully — matters.

When AI Is Used Outside Its Lane

The research above says psychotherapy is a low-exposure job — meaning AI is not about to replace trained therapists. But that finding is being misread by the public.

A growing trend has teenagers using ChatGPT, Claude, Gemini, and Character.ai chatbots as informal therapists. The 2025 research on this practice is not optimistic.

My takeaway: "AI cannot replace therapists" applies to clinical work done by trained, accountable humans. It is not a green light for teens to confide in chatbots. General-purpose models are tuned to be fluent and agreeable — not to detect crises, push back gently, or hold confidentiality. That gap matters most when the user is most vulnerable. The right use of AI in mental health is as a tool for clinicians and a triage front-door — not as a substitute friend or therapist for a 14-year-old at 2 a.m.

Test What You Know

Eight questions drawn from the data above. Immediate feedback per question.

Sources

  1. IMF: AI Will Transform the Global Economy (2024)
  2. WEF Future of Jobs Report 2025
  3. Stanford AI Index 2025
  4. ILO: Generative AI Likely to Augment Rather Than Destroy Jobs
  5. McKinsey: The Economic Potential of Generative AI (2023)
  6. Acemoglu: The Simple Macroeconomics of AI (NBER, 2024)
  7. Stanford HAI: Exploring the Dangers of AI in Mental Health Care (Moore et al., FAccT 2025)
  8. Common Sense Media & NORC: Talk, Trust, and Trade-Offs — How and Why Teens Use AI Companions (2025)
  9. APA: Using generic AI chatbots for mental health support — A dangerous trend (2025)
  10. APA Health Advisory: AI & wellness apps alone cannot solve mental-health crisis (Nov 2025)