EXPOSED
Lecture content, problem sets, exam writing, and even first-pass grading of stoichiometry and mechanism questions are the kind of structured text work AI already does at usable quality, and open courseware plus tutoring bots erode the value of the standard general-chemistry lecture. What does not transfer is the wet lab: supervising students handling reagents, enforcing hazard protocol, diagnosing why a titration or synthesis failed, and running a research group. The modal worker here is heavily weighted toward large intro-course teaching and adjunct or non-tenure appointments, where enrollment economics — not AI capability — is the nearer threat.
Mixed — a routine tier and a judgment tier. Writing the general-chemistry syllabus, generating stoichiometry and equilibrium problem sets with worked solutions, and marking multiple-choice or numeric-answer exams are already handled at usable quality by current tools, which caps this below the mixed range's midpoint; what pulls it up to 12 rather than 6 is the unscriptable half of the job — standing at a fume hood judging whether a student's setup is about to boil over, and reading a failed IR spectrum with the student to find where they went wrong.
Hands-on in uncontrolled environments. A 13 reflects that the teaching lab is a genuinely uncontrolled environment — open flames, concentrated acids, unpredictable undergraduates — and the instructor is physically present enforcing goggles, waste segregation, and fume-hood use, not just certifying that a procedure exists on paper; it stops at 13 rather than 18 because the lecture, office hours, and grading portion of the load is entirely screen-and-chalkboard work.
No licence, no signature requirement. There is no state license to teach chemistry at a college — a PhD and a departmental appointment are the whole gate — so despite bearing real OSHA lab-safety and chemical-hygiene-plan duties, nothing personally attaches to you the way a PE stamp or medical license does, and the institution absorbs the exposure; the 3 rather than 0 is that hazardous-materials responsibility, which is delegated to you by name in most chemical hygiene plans.
The human relationship is the product. At 14, this recognises that research-group mentorship, thesis advising, and the letters of recommendation that decide whether a student gets into a PhD program are irreplaceable relationship goods — a bot cannot vouch for someone's bench technique; it is not higher because the modal enrollee in a 300-seat intro section is interchangeable to you and you to them, and adjunct appointments break the multi-year continuity that would push this to 18.
Meaningful discretion. Deciding whether a student's inconsistent yields are sloppy technique or a contaminated reagent lot, setting the curve on a failing exam, and calling whether an unfamiliar reaction is safe to let sophomores run are real discretionary calls, but they sit at 12 rather than 16 because course learning outcomes, ACS-aligned curricula, accreditation requirements, and departmental grade norms already define the boundaries you are choosing within.
Physician Assistants SAFE
Nurse Anesthetists SAFE
Anesthesiologists SAFE
Has AI actually changed your work?