Moderate pressure on university lecturers. The shape is what matters.

University lecturing is the role most likely to move sideways rather than disappear. Students are using LLMs daily; the smarter universities have already restructured assessments to assume that.

Automation is medium-low and rising. Recorded explainer content, literature-review drafts, routine assessment-generation, marking standardised work. Big lecture courses with multiple-choice exams will look very different in 2030 — closer to a curated playlist with model-graded checkpoints than a hall of 400.

Augmentation is high. Research synthesis across long bodies of work is the most legitimate use of AI in academia today, and it's saving lecturers and PhD candidates real time. Tutorial prep, language adaptation, curriculum updates from the latest field results — all faster.

What stays human: postgrad mentorship, original research direction, examination of hard cases, viva voce, grants. The lecturers most under pressure aren't the research-active ones; they're the ones whose value was "I deliver the lecture you can also watch on YouTube." The pivot path is up the research ladder, not sideways.