Example · PhD, Life Sciences

88% of the work a PhD student in life sciences is heading toward still needs a person.

Grant funding is tighter than it was. Your bench work and your research stay with you.

Make my planA typical week at the middle pace. Your plan uses your own tasks.

Is a life sciences PhD worth it with AI?

In a typical week at the middle pace, 58% of this work needs your judgment and 30% is work you do using AI. AI could do about 12%. Grant math moved more than the models did. The success rate at NIH, the National Institutes of Health, for research project grants fell to 13% in fiscal 2025, down from 18.5% a year earlier (NIH Extramural Nexus, March 2026).

Read more

Meanwhile the parts of your PhD that need your hands at the bench and your name on the paper stayed hard to hand to software. Most of your week is that lasting work. The exposed slice is the literature digging, the routine analysis and the slide decks, and it is exposed fast.

That makes your job for the next year fairly specific. Become the person in your lab who knows exactly where a model's analysis or literature summary goes wrong, and write it down. That record helps in industry interviews, in a fellowship application and in your next committee meeting. If committee meetings are hard for you, the fix is to walk in with the agenda and a decision you want them to make. You don't need to leave the bench, change fields or become a programmer.

The week, task by task

Who does each hour of a typical 40-hour week.

See every task
You58% · 23 hours

Needs your judgment.

  • Bench experiments: cloning, cell culture, assays, sample prep19%
  • Teaching sections, grading, and mentoring rotation and undergraduate students10%
  • Drafting manuscripts and dissertation chapters8%
  • Animal work and specialized instruments under an approved protocol7%
  • + 3 smaller tasks14%
You using AI30% · 12 hours

AI does the slow parts. You still own it.

  • Analyzing data: statistics, plots and analysis scripts10%
  • Image analysis: segmentation, counting and quantifying microscopy6%
  • Designing protocols and planning the next experiment6%
  • Lab operations: ordering, inventory, instrument booking and logs5%
  • + 1 smaller task3%
AI12% · 5 hours

AI could do most of it.

  • Literature searches and background synthesis for papers, quals and lab meeting7%
  • Industry: project-team reports, slide decks and data summaries5%
Why some tasks still need a person
Cloning, culture, assays and sample prep at the bench

Autonomous labs and university cloud labs exist. Most academic biology still happens by hand, and every automated run still needs someone to choose the question and catch the bad plate.

Animal work under an approved protocol

Federal rules make the research facility ensure that "all scientists, research technicians, animal technicians, and other personnel involved in animal care, treatment, and use are qualified to perform their duties" (9 CFR 2.32(a)). Your training record is part of what lets the experiment happen.

Finding out why an experiment failed

A model can list twenty causes, but you know the incubator ran warm last week and the antibody lot changed.

Your name on the paper

ICMJE says chatbots "should not be listed as authors" and "humans are responsible for any submitted material that included the use of AI-assisted technologies."

Mentoring a rotation student, running a TA section

Students come back to the person who helped them, and hiring committees ask about it.

Good to know

What is hard about this field, and what you can do about it.

The hard truth

A postdoc pays half what industry pays new PhDs.

Years of postdoc pay add up to a large cost. Make it a choice with a clear purpose.

What you can doTake a postdoc only for a goal it serves, like a faculty path or a specific skill. Build industry contacts, internships and transferable skills during your PhD.
Sources (1)

What careers come after a life sciences PhD?

Plan A is the best move in the field you are studying for.

APlan A

Focus your studies

The main focus. Most weekly sessions build toward it.

Your next stepAim for industry research scientist jobs in biotech or pharma
Each week
4 sessions
Ready in
One semester for the evidence and the resume
Why it fits

Why you. You already do the bench work and the analysis. Industry teams want people who can do both and say which model output to trust.

The gap. You have no evidence of checking AI outputs, no industry contacts, and an academic CV instead of a two-page resume.

The trade-off. Biotech layoffs come in waves: 62 announcements in Q3 2025 and 21 in Q3 2026 (Fierce Biotech). The market is calmer now, but that could change next year.

2 other moves that fit
  • Take a postdoc for a clear reason, such as a spot in NIH's own labsReady in nine to twelve months
  • Look at jobs next to research: medical science liaison, regulatory affairs or science policyReady in one to two years after the phd

What's a plan B for biology PhDs?

Plan B is meant to be different from this job. It was picked because skills from this role carry over.

BPlan B

Get ready for a new kind of job

A second option. A newer job AI is creating that these skills already fit.

A job to aim forExpert AI trainer (contract)

Finding out why an experiment failed is the reasoning expert platforms pay life science PhDs to apply when grading AI answers.

Each week
1 session
Outlook
Hiring now
Pay
$50 to $110 an hour (contract)Three legal AI trainer contracts in 2026; the top rate requires a JD.
2 other options
  • AI policy analystScience policy fellowships suit PhDs, and standing in front of your committee prepares you to brief decision makers on AI in biology.
  • Domain AI red teamerYour bench and protocol knowledge lets you probe where AI models give unsafe or wrong lab guidance, a growing area of red teaming.

In your own plan you can switch to any of the 47 roles.

The 13 weeks

Small steps that add up. One plan for both: most sessions build Plan A, one a week builds Plan B.

By week 13 you will have

AFor Plan A

  • A validation log of where AI tools went wrong in your field
  • A two-page industry resume and at least one submitted internship application

BAnd for Plan B

  • A timed practice record
  • An expert task pack for your portfolio and platform applications
Month 1Weeks 1 to 4
  1. 1See where your time goes
  2. 2Start your validation log
  3. 3Test image and structure tools
  4. 4Check a literature summary

You finish withA validation log with about 30 entries

Month 2Weeks 5 to 8
  1. 5Decide on the F31
  2. 6Thanksgiving: a lighter week
  3. 7Talk to industry
  4. 8Grading week: a lighter week

You finish withAn up-to-date log and resume

Month 3Weeks 9 to 13
  1. 9Send a committee pre-read
  2. 10Holiday: build your tracker
  3. 11Holiday: draft your aims
  4. 12Get feedback
  5. 13Submit applications

You finish withA second application or three postdoc inquiries

Three courses from the plan
  • Parent F31 Individual Predoctoral Fellowship (PA-25-422)NIH · Free to apply; predoctoral stipend $29,364 in FY2026

    Yes, Week 5.

  • NSF Graduate Research Fellowship (NSF 26-526)NSF · $37,000 stipend plus $16,000 to your school

    No, if you are in your second year or later. Tell the first-years.

  • NIH policy on AI in applications (NOT-OD-25-132)NIH · Free

    Yes, before you draft anything.

Checked 2026-10-13.

The fuller picture

Risks beyond AI: highGrant and exit risk

Grant policy sets how many stipend and postdoc lines exist. NIH is making fewer new research grants even with a steady budget, and new life sciences PhDs had the lowest share of definite job commitments of any field in 2024. If you are an international student, a proposed OPT fee is open for comment and is not in effect.

Where this could be wrong
  • The shares assume a wet-lab student with animal work and imaging. A computational or structural biology student spends far more time on analysis, and much more of their week is easy to automate.
  • Bench automation may move faster than we scored. If your institution opens a cloud lab or your lab buys a liquid handler, automated tools start assisting with bench work.
  • The Survey of Earned Doctorates numbers describe 2024 graduates. One cohort is a weak trend, and the 2025 data could reverse the industry dip.
  • The Kosmos accuracy figures come from its developers' own evaluation of 102 statements. Independent replication is limited.

This is an example. Your plan uses your own week.

Make my plan

Also matched: phd student biology, biology phd, grad student life sciences. Research checked 2026-10-13. Page updated .