No, AI can’t be your teaching assistant 

Artificial intelligence is in McMaster’s classrooms and taking tasks away from its teaching assistants

By: Kate Linardic, Opinion Contributor

Artificial intelligence is beginning to integrate itself into life at McMaster and recently it has encroached upon our pipettes and Erlenmeyer flasks. Stemble is an AI grading software being used by the Department of Chemistry and Chemical Biology in its Level I courses to ease the burden of large class sizes.

Once the responsibility of paid teaching assistants, Stemble has absorbed the job of marking post-lab assignments and homework in Level I chemistry courses. The ability of AI models like Stemble to take on more responsibility is only increasing. 

Level I chemistry midterms and final examinations have long been graded using scantrons. This was done to offload the work of grading, leaving teaching assistants with the responsibilities of attendance-taking, laboratory supervision and assignment-marking. 

But can AI fully replace TAs? The answer is complicated. 

AI is being rapidly welcomed into workplaces and schools while regulations are struggling to keep up. Its appeal is the opportunity it provides employers to reduce labour costs. 

In academia, services like Stemble enable departments to cut back on the number of hours spent grading in TA job descriptions. Meanwhile, the cost of Stemble pops up as a separate subscription fee to be paid for by students. 

Despite the benefits of AI grading for the university, there is an element of humanity innate to teaching that is not quite yet able to be replicated by AI.  

A 2023 study at the University of Queensland found that 55 per cent of its global respondents were not yet comfortable with AI taking on people-management roles, like that of supervising undergraduate students performing titrations in the A.N. Bourns Building chemistry labs. 

In its 2024 report, the World Economic Forum found that it is unlikely that current AI developments will impact the interpersonal aspects of teaching or replace the need for direct interactions between teaching assistants and students. It appears that a person-to-person connection adds something distinctly valuable to the learning process and remains a practice which still requires both participants to in fact be persons. 

With these aspects considered, AI is unlikely to result in major job losses for the time being, meaning chemistry TAs can rest assured that they have a few more years of answering the same questions about where students can put their glassware at the end of a lab. 

However, this may only hold true for the time being. The AI industry is only starting to pick up steam and new technology is developed every day. But even if AI’s teaching capacities improve, it cannot be a role model for students pursuing academia the way that student-TAs can. 

TA positions also give students a critical opportunity to try teaching and mentoring, something that those pursuing academia will possibly be expected to engage in depending on where they find work. It serves as a trial run for those who may see themselves as future professors and instructors. If TA positions were to disappear, so would the opportunities for students to gain experience in teaching and mentoring others. 

Student TAs have an intrinsically human element to offer undergraduate students, unlike their environmentally damaging AI counterparts. And yes, AI may soon develop such “human” abilities, too, but there is no need to rush into that future when the alternative is continuing to grant real chemistry students an opportunity to share their knowledge with others. 

McMaster should continue to take care when investing in AI like Stemble. In an increasingly atomized world, keeping open avenues for communal learning at McMaster should be regarded as important as ever.