TCD · The Career Diagnostic

Will AI take my chemist job?

Science · SOC 19-2031
73
/ 100
Career Diagnostic Score · science

Chemist

In 1 year
71
↓ 2 pts
In 3 years
68
↓ 5 pts
In 5 years
65
↓ 8 pts

Tasks at risk

  • Standard literature reviews
  • Routine synthesis-procedure drafting
  • Spectra interpretation for common compounds

Skills to build that AI can't replace

  • Bench experimentation and lab safety
  • Novel synthesis and discovery work
  • Scale-up and process chemistry

Get your free 12-week career plan for Chemist

A research-backed plan tailored to your job: which AI tools to learn, which skills to deepen, which credentials just got more valuable. Delivered to your inbox.

Free · No spam · Unsubscribe in one click · how the score is computed

What this score means for Chemist.

The Chemist role is durable. AI is squeezing the documentation surface and some routine knowledge work, but the value-creating core remains squarely human. The right move is to use AI to take the paperwork off your plate so you can spend more time on what AI cannot do. Over the next 5 years the score is projected to drift down by 8 points based on the science vertical's drift rate. We update the score every 3 months as new data comes in.

The four signals.

Can AI do it directly?
10%
Share of tasks AI can already do on its own.
Could AI tools do it soon?
45%
Share of tasks AI tools could plausibly handle soon.
Are people using AI for this now?
30/100
How much real workers are using AI for this work today.
Replacing or helping?
22%
Of that usage, how much is AI replacing vs. helping the worker.

All four are blended into your single 0–100 score. See the exact formula →

What to do this quarter.

The single highest-leverage move for someone in this role this quarter:

A frontier AI assistant for literature review and protocol drafting; humans for the bench work.

The free 12-week email plan above expands this into a sequenced plan. The $29 Pivot Interview goes much further: ranked tools, credential-value shifts, pivot scenarios, and quarter-by-quarter targets for the next 12 months.

Other roles in Science.

Sources for this score.

  • Anthropic Economic Index — Anthropic's real usage data (March 2026 release), used for the "real usage" and "replacing vs. helping" signals. huggingface.co/datasets/Anthropic/EconomicIndex
  • Eloundou et al. (2023) — the Stanford / OpenAI study used for the "can AI do it" and "could AI tools do it soon" signals. arXiv:2303.10130
  • U.S. Department of Labor (O*NET 28.0) — the official task-and-skill list for every job in America. onetcenter.org/database
  • U.S. Bureau of Labor Statistics — the official jobs classification (SOC code). bls.gov/oes