Teaching in the age of ChatGPT: an operator’s view

I’ve been on the front lines of adjunct instruction for years: part-time Earth science faculty work, combined with other jobs, driven more by the satisfaction of teaching than by pay or stability. That intrinsic reward can be powerful — until external changes make the job harder. In this account, one instructor explains how the arrival of generative AI has turned a previously rewarding role into a largely miserable one in certain contexts.

The specific setting that amplifies the problem is asynchronous, recorded instruction. Over the last few years this instructor has taught exclusively in that format, and the combination of remote, on-demand delivery and new AI tools has exposed structural weaknesses in how learning and accountability are designed.

Why instructors accept recurring friction

There are three facts that matter for understanding the baseline: this instructor is part-time, teaching Earth science; teaching is done alongside other paid work; and the motivation to teach is primarily the human reward of working with students. These conditions are common in higher education and explain why many talented people accept low pay and tenuous employment.

The carrot is human interaction. Despite financial and contractual downsides, the instructor continues because student engagement is fulfilling in a way that other roles are not. That satisfaction is the key retention lever for adjunct faculty.

Asynchronous delivery: the operational disadvantages

Recorded, asynchronous courses change the signal available to an instructor. Where a live classroom gives immediate, visible cues — attendance, facial expressions, real-time questions — the recorded format removes that feedback loop.

Mechanics of degraded oversight

• Students don’t have to be present at a fixed time, so monitoring engagement becomes indirect.
• Instructors lose spontaneous checks for comprehension that live interaction provides.
• Without visibility, the probability that students disengage or stop participating rises.

Put plainly: asynchronous designs increase the risk that students will "fall off" because there is no enforced, shared moment of attention where an instructor can steer and intervene.

Generative AI as an aggravator, not a root cause

The instructor’s assessment is stark: generative AI has shifted the experience from challenging to mostly miserable in certain settings. It is important to read that as a symptom interaction rather than a standalone diagnosis.

When the course is recorded and students can work on their own schedule, new tools amplify the existing weaknesses of the format.

The claim is not that AI single-handedly destroyed teaching; instead, the combination of a format that lacks live accountability and platforms that change how students access answers has created new frictions for instructors who rely on student interaction for motivation and for course functioning.

Practical levers that follow from this account

We are working from a single, consistent narrative. Still, it yields operationally useful directions for teams designing courses or programs where part-time instructors and asynchronous delivery are core components.

• Reintroduce synchronous touchpoints. If recorded material must exist, pair it with mandatory, short live checkpoints to restore visibility and quick feedback loops.
• Make participation observable. Design low-friction evidence of engagement (time-stamped reflections, small project milestones) so instructors can detect disengagement early.
• Preserve the human reward. Create structured moments where instructors interact with students in ways that are intrinsically rewarding — critique sessions, office hours, peer review — so teaching remains sustainable for adjuncts.
• Accept format trade-offs. If the program relies on asynchronous scale, acknowledge the operational cost: higher risk of drop-off, greater need for automated signals, and potential instructor dissatisfaction.

Where this fails for faculty

For an instructor compensated poorly and balancing multiple jobs, the extra labor required to retrofit asynchronous courses with accountability measures changes the work calculus. The reward of human interaction diminishes if most time is spent chasing signals rather than teaching. That’s the core grievance reported: the job stops being fulfilling when structural changes make the emotional payoff scarce and effort high.

What This Means For You

If you run programs that use part-time faculty or scale via recorded content, treat this account as an operational red flag rather than a moral panic. The combination of asynchronous delivery and external tools that alter how students engage will surface latent weaknesses quickly.

Concrete, immediate actions to consider:

• Map where human feedback is essential. Prioritize synchronous or visible checkpoints for those moments.
• Design participation metrics that are cheap to collect and meaningful to instructors.
• Compensate instructors for the extra coordination work required to maintain engagement in asynchronous formats.
• Experiment with short pilot changes (one mandatory live session per module, or weekly micro-assignments) and measure instructor satisfaction as a leading indicator.