EXPOSED
Lecture drafting, slide decks, syllabus writing, rubric-based grading of essays and speech reflections, and discussion-prompt generation are already handled competently by LLMs — that's a large share of the weekly workload. What holds is the live room: coaching a nervous student through a five-minute persuasive speech, reading an audience in real time, running peer critique, and being the person whose judgment on a student's argument actually counts. The bigger threat isn't AI replacing the professor, it's enrollment decline plus adjunctification compressing headcount while AI absorbs prep and grading labor.
Mixed — a routine tier and a judgment tier. Half the job — writing the public-speaking syllabus, building rhetoric lecture decks, generating discussion prompts on media effects, marking speech self-reflections against a rubric — is text-in/text-out work an LLM does at passable quality, while the irreducible half is standing in a room watching a student's hands shake through their informative speech and deciding in real time whether to interrupt or let them finish.
Some physical or field component. You are physically in a classroom or speech lab most contact hours, positioning a camera, watching posture and eye contact and vocal projection in a body you can only assess in person — but it's a scheduled, climate-controlled room on a campus, not a job site, which is why this sits mid-scale rather than high.
No licence, no signature requirement. No state licence gates who teaches COMM 101 — a master's degree and a department chair's hiring decision is the whole barrier, and adjuncts are hired on that basis every August with no personal legal exposure for what happens in the room.
The human relationship is the product. Students take Public Speaking terrified, and the reason they show up and actually deliver is that a specific instructor learned their name, remembered last week's speech, and gave feedback they trusted; that relationship also produces the recommendation letters and mentorship that survive long past the course, though the intro-course, fungible-section structure keeps it below the 17-20 range of true one-to-one practice.
Meaningful discretion. You decide whether a student's argument on a contested political topic gets protected as advocacy or flagged as a classroom-conduct issue, whether a plagiarized or AI-drafted speech outline becomes an academic-integrity referral, and where the line is on a C+ versus B- for persuasive delivery — real discretion, but exercised inside departmental rubrics, appeal processes, and accreditation-driven learning outcomes that constrain the call.
Has AI actually changed your work?