Example · PhD, Life Sciences
Your bench work and your byline still need you. The bigger change is tighter grant money.
35 of your 40 hours still need you
- 23 stay yours
- 12 faster with AI
- 5 tools could take
A typical week at the middle pace. Your plan uses your own tasks.
Is a life sciences PhD worth it with AI?
In the work a PhD student in life sciences is heading toward, 35 of a typical 40-hour week still need a person, by our task-by-task count. Tools could already take about 5 hours. 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 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.
Sources (1)
- In 2024, new biological and biomedical PhDs expected a median $60,000 in postdocs versus $120,000 in industry. NSF NCSES, Survey of Earned Doctorates, 2026-02
What stays yours
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.
What careers come after a life sciences PhD?
Plan A is your best move in the field you are studying for. Most sessions in the 13-week plan build toward it.
- A
Aim for industry research scientist jobs in biotech or pharma
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.
Other moves that fit
These are also good bets in this field. Read them if Plan A doesn't suit you.
Take a postdoc for a clear reason, such as a spot in NIH's own labs
Why it fits
Why you. If you want a faculty job, or a skill your lab can't teach, this is the road.
The gap. You need a clear reason and a specific lab, plus the habit of asking about funding before you accept.
The trade-off. This move is the most exposed to grant policy. A first-year NRSA (NIH training award) postdoc stipend is $63,480 in FY2026 (NIH NOT-OD-26-044), and institutional salaries vary.
Look at jobs next to research: medical science liaison, regulatory affairs or science policy
Why it fits
Why you. This fits if you like explaining science more than doing it, or if you found the IACUC (animal care committee) and biosafety work more interesting than you expected.
The gap. These jobs usually want experience you can't get in the lab. The RAC exam, the regulatory affairs certification from RAPS, the regulatory professionals' society, requires at least one year of regulatory-related experience with a doctorate. Policy fellowships such as AAAS's require a completed PhD, and some placements require US citizenship.
The trade-off. You leave the bench. Medical science liaison hiring data are thin and mostly from recruiters.
What's a plan B for biology PhDs?
Plan B is a job AI is creating that these skills already fit. One session a week in the plan trains you for it, so it is ready if your job shrinks or you want a change.
BPlan B, picked for this role
Expert AI trainer (contract)
Finding out why an experiment failed is the reasoning expert platforms pay life science PhDs to apply when grading AI answers.
Other options
AI policy analyst
Science policy fellowships suit PhDs, and standing in front of your committee prepares you to brief decision makers on AI in biology.
Domain AI red teamer
Your 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 first two weeks
The first 2 of 13 weeks. Most weeks have three sessions of 60, 30 and 90 minutes, plus one 40-minute Plan B session (45 minutes in week 4). The plan assumes a mid-October start, so weeks 6, 8, 10 and 11 have one session plus the Plan B session for Thanksgiving, grading week and the holidays. By week 13 you'll have a validation log of where AI tools went wrong in your field, a two-page industry resume and at least one submitted internship application.
- Week 1: See where your time goes
What you do
- AList a typical week of your bench work, classes, teaching and writing. Mark which parts AI already helps with.
- ARead the biology row of the Survey of Earned Doctorates, the national survey of new PhDs. Then read NIH's FY2025 grant numbers and write one paragraph on where you stand.
- ATake the myIDP assessment, a free career planning tool for scientists.
- BTake the first half of Anthropic's AI Fluency: Framework & Foundations course.
- Week 2: Start your validation log
What you do
- AAsk your PI, the head of your lab, for the lab's rule on AI tools. Ask for it in writing.
- APick a public dataset from a published paper in your field.
- AHave a model analyze the dataset. Check every step against the paper and record each problem in a validation log.
- BFinish AI Fluency. Ask AI five hard questions from your subfield and grade each answer against published papers.
Three courses from the plan
Checked 2026-10-13.
- 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.
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.
Also matched: phd student biology, biology phd, grad student life sciences. Research checked 2026-10-13. Page updated .