Why this job stays with people
Will AI replace archeology teachers? Not in one move. The job bundles three different kinds of work, and they pull in different directions. There is live teaching. There is original research and fieldwork. And there is the department work that keeps a degree program running: advising, committees, accreditation paperwork, grant writing.
Language models are good at the text layer of that bundle. Drafting a lecture outline on Mesoamerican ceramics, building a reading list, writing a first pass at feedback on a term paper, summarizing a new site report. Those are real tasks, and they take real hours. But the hours that define the job are harder to hand over. Supervising a field school, where a student has just cut through a feature and has to decide what to do next. Teaching lab method by watching hands. Judging whether a graduate student is ready to defend. Sitting with someone who is thinking about leaving the program.
Scale matters too. The Bureau of Labor Statistics counts about 5,240 people in this occupation, with median pay of $99,650 and projected employment change of 2.6% from 2025 to 2035 (BLS, 2025). This is a small field. Enrollment, state budgets and tenure-line decisions move it more than any tool does. Where AI shows up first is in how many hours a course takes to run, not in whether the course exists.
What AI does, what it helps with, what it leaves alone
Tasks where a model can take the lead account for 2% of task time here. These are the routine text jobs around a course: turning a syllabus into weekly materials, producing quiz banks and reading questions, converting lecture notes into slides and handouts, and first-pass summaries of literature for a seminar. The output still gets checked. A model will happily invent a citation or flatten a contested interpretation into a tidy consensus.
Assisted work is the bigger slice, at 56%. Grading sits here: a tool can sort and pre-mark structured answers, but the comment that changes how a student argues is still written by the instructor. So does research support. Machine learning already does serious work in archeology itself, including classifying ceramic sherds and lidar imagery, clustering survey data, and drafting sections of grant proposals and papers. The teacher directs it, checks it and signs their name to the result.
What stays with people is 42% of task time. Field and lab supervision, where safety and irreversible decisions are in play. Student advising and mentoring over years. Curriculum design and the judgment calls behind a program. Collaboration with descendant communities, museums and permitting bodies, where trust and accountability sit with a named person. The task list above shows which duties land in each group.
What the evidence actually shows
There is no direct head-to-head test of AI against postsecondary anthropology and archeology teachers. That is why the parity grade here is D, and why no parity number is given. A grade at that level means the quality question has not been measured for this job, not that AI did badly. We publish the rule in the quality parity method.
What would settle it is specific and doable: blind comparison of AI-written and instructor-written feedback on graded student work in the discipline; measured learning outcomes in sections taught with and without AI support; and accuracy testing on domain tasks such as artifact classification and site-report synthesis, scored by specialists. Until that exists, the evidence list on this page is what there is, and the general-purpose studies are a weak proxy for a field that runs on physical material.
Coverage is a separate question from quality. It estimates the share of task time AI can handle today, and for this job it reads 37 out of 100. The reasoning behind that figure is in the coverage method, and the whole scoring approach is set out in our methodology.
When the picture could shift
Most likely between 2034 and 2046 (8 in 10 of our scenarios). How that window is built is explained on the replacement year page.
Two things could pull it earlier. First, budget pressure: the cost of running a model across a semester of course admin is far below a teaching line, so a department under strain may cut sections and raise class sizes rather than cut tools. Second, fewer junior openings. Adjunct and visiting posts are where routine teaching load lives, and that is the load tools reduce first.
Two things hold it back. Accreditation and tenure rules put a named instructor behind every course and every grade, and that is slow to change. And the core of the discipline is physical: excavation, curation, condition assessment, handling fragile material under permit. Only a small share of the job is physical work, so hardware is not the brake here. The brake is accountability and the field itself.
Good to know: AI in this occupation is showing up as task erosion inside existing posts, not as departments teaching without faculty.
How to stay needed
Lean into the work that sits in the needs-a-human group. Run field schools and lab practicums, and own the safety and method training that goes with them. Take advising and thesis supervision seriously, and keep records of where students land. Lead curriculum and assessment design, including the policy on what AI use is allowed in your courses.
Two skills pay off. One is practical fluency with machine learning in archeological analysis, enough to supervise a student using image classification or survey clustering and to catch a bad result. The other is public and community work: writing, museum collaboration and consultation with descendant communities, which is judged on trust rather than output volume.
Close jobs are worth comparing. Look at history teachers, postsecondary, sociology teachers, postsecondary and anthropologists and archeologists, the practitioner side of the same field. You can also see the whole postsecondary teachers family, the wider education sector, or the jobs that most need a person list. To put two of them side by side, use the job comparison tool.