EXPOSED
Lecture slides, syllabi, problem sets, rubric-based grading of lab reports, and literature reviews for grant proposals are all things current models draft at usable quality, and enrollment pressure in a small field (6,690 jobs) makes course consolidation into shared or asynchronous sections a real threat. What holds is the embodied and accountable part: running field sampling trips and wet labs, mentoring graduate researchers, sitting on IRB/thesis committees, and signing off on data and publications under your own name. The modal worker here is a tenure-track or contingent faculty member whose survival depends more on institutional budgets and enrollment than on model capability.
Dipped in 2020, then grew past where it started.
Median pay $82,430 → $94,980 -7.8% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
+2.9%
Percentage only. The projection counts a different population from the 6,690 above — it includes self-employed workers, which for this occupation is most of them, so the two headcounts are not comparable.
Growing, and only partly exposed
The BLS expects +2.9% more of these jobs by 2034, and at 53/100 the work is only partly exposed — some tasks are automatable, the core of the job is not. Nothing here is in tension.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~700 openings a year on average, including replacing people who leave.
EducatorLecturerProfessorInstructorFaculty MemberAdjunct ProfessorCollege ProfessorAssistant ProfessorAssociate ProfessorConservation EducatorCollege Faculty MemberEnvironmental EducatorUniversity Faculty MemberNatural Resources ProfessorWater Conservation EducatorConservation Science TeacherEnergy Conservation EducatorUrban Environmental EducatorEnvironmental Science ProfessorEnvironmental Studies ProfessorEnvironmental Engineering ProfessorEnvironmental Studies Faculty MemberEnvironmental Science Management and Policy Professor
Holding it up: trust premium . Weakest point: liability shield .
Mixed — a routine tier and a judgment tier Roughly half the job — building slide decks on biogeochemical cycles, writing exam banks, grading lab reports against a rubric, and summarizing recent literature for a seminar — is already draftable, but supervising a stream gauging exercise, calibrating a YSI sonde with twenty undergrads in waders, and reading a struggling thesis student's raw data for what they actually did wrong are not, which is what puts this at 11 rather than 6.
Some physical or field component Field methods courses, watershed sampling trips, GC/ICP-MS and wet-chem lab supervision, and the safety judgment calls that go with cold water, ticks, and hazardous reagents are genuinely physical, but the majority of contact hours are still lecture, office hours, and committee rooms — an 11 reflects a real field component that does not dominate the teaching load.
No licence, no signature requirement There is no licence to teach environmental science; a PhD is a hiring credential, not a statutory gate, and while tenure and accreditation reviews create institutional friction, no board can strip your right to practice, so the 4 comes from the accreditation/HR layer rather than any personal licensure.
Meaningful discretion You decide whether a thesis defends, whether a student's field data are clean enough to publish, whether a site is safe to sample, and how to handle authorship disputes and academic-integrity cases — high-stakes, ambiguous, and yours to own — with 13 rather than 16 because departmental curriculum committees, IRB, and deans formally share the heaviest of those calls.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (11/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (4/20) is whether the law requires a licensed human to sign. Trust premium (14/20) is whether buyers specifically pay for a person. Judgment and accountability (13/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 31 of this occupation's 53 points (58%).
Embodiment (11/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 65/100, still EXPOSED.
Task-mix shift: once slide decks, problem sets, and rubric grading are conceded to models, the remaining tier is grad-student research mentoring, field method design, and thesis committee work. The field genuinely has these two tiers, so if institutions formally reassign the routine tier to shared/AI-assisted asynchronous sections while keeping named faculty on research supervision, the residual job is the judgment tier
Named-author and data-integrity obligations tightening: NSF/NIH research-misconduct rules and journal policies (e.g. ICMJE-style authorship, Nature/Science AI-authorship bans) put the PI personally on the hook for data provenance and for whether AI-generated text or analysis entered a submission. Institutional research-integrity offices adding AI-disclosure attestations signed by the supervising faculty member raises this further
Accreditation or program-review language that ties environmental science degree approval to a minimum number of supervised field/wet-lab contact hours (as ABET-accredited environmental engineering programs already specify laboratory experience) — making the instructor-led field sampling trip, boat/electrofishing safety supervision, and instrument-calibration lab non-substitutable by asynchronous delivery
Field-course risk management: campus risk-management and insurer requirements for off-campus field trips (wilderness first aid certification, named trip leader on the liability waiver, water-safety credentials) create a documented individual signatory for student safety. Narrow, and only applies to faculty who actually run field courses
State licensure pipelines: if more states follow the pattern of licensed Professional Geologist / environmental professional boards in requiring coursework taught by qualified faculty at accredited institutions for exam eligibility, employers and students pay for the credential chain rather than the content — but this is weak for environmental science specifically, which lacks a single dominant license
The limit. The binding constraint here is not model capability but enrollment and budget. A 6,690-job field faces course consolidation and contingent-hiring decisions made by deans, and no liability or accountability lever protects a section that gets merged or an adjunct line that isn't renewed. Realistic ceiling is low-to-mid 60s, and even that mostly protects tenured faculty who run field programs and supervise graduate researchers — not the contingent lecturer teaching intro sections.
| New York-Newark-Jersey City, NY-NJ | 530 | $103,460 +9% |
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 200 | $104,810 +10% |
| Atlanta-Sandy Springs-Roswell, GA | 190 | $110,540 +16% |
| Boston-Cambridge-Newton, MA-NH | 170 | $107,010 +13% |
| Austin-Round Rock-San Marcos, TX | 140 | $77,190 -19% |
| Portland-Vancouver-Hillsboro, OR-WA | 120 | $85,740 -10% |
| Seattle-Tacoma-Bellevue, WA | 110 | $130,260 +37% |
| Chicago-Naperville-Elgin, IL-IN | 100 | $90,030 -5% |
| Seattle-Tacoma-Bellevue, WA | 110 | $130,260 +37% |
| Minneapolis-St. Paul-Bloomington, MN-WI | 60 | $127,100 +34% |
| Atlanta-Sandy Springs-Roswell, GA | 190 | $110,540 +16% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 53. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.