Although it seems to be the argumentative equivalent of spilling a glass of water into the Pacific with the goal of flooding Sydney, I’m voicing my concern for the humanities in the ever-expanding face of artificial intelligence. The arguments against AI’s encroachment in academic settings, though prolific, have done nothing to mitigate it. A similar source of adversity facing English departments in particular, is the growing inability of college students to read long-form literature. Note my usage of the word inability; students are not expressing boredom or a lack of time in response to being assigned novels, but rather a complete inability to read them.
With the increasing presence of AI and its ability to both summarize and analyze literature, combined with the simultaneous abetment of students’ attention spans, there exists the danger of a positive feedback loop to propel students down a steep spiral of ineptitude. AI, becoming more advanced in nature each day, supplements students’ decreasing attention spans by giving them the tools to supplant their readings. Thus, the coffin of students reading long-term literature seems all but nailed. There is, however, a mutually beneficial solution that can be feasibly achieved should anti-AI purists raise their heads out of the sand and face the artificially generated music.
As it stands, the syllabi in my English classes generally say the following in regards to AI use: Students are not allowed to use any generative artificial intelligence tools, e.g. ChatGPT, at any stages of the work process, including preliminary ones. The message is clear: AI, in any capacity, is forbidden. By extension, the use of AI in any fashion, even one that doesn’t necessarily replace individual thought and analysis of texts, is cheating.
This stipulation ignores the many areas where AI can prove ultimately beneficial. One domain that I have found generative AI to be useful is grammar. Google Docs’ spelling and grammar check only scans for blatant violations of standard grammar rules and offers direct replacements. This service, which is permitted for use, only fosters learning in the field of pressing buttons and using Google Docs.
More complex aspects of grammar tend to fall under the radar. One specific area to which I pay particular attention is the passive and active voice, something that is not typically flagged by grammar checks. In instances where AI is not completely disallowed, services that read my paper and highlight areas of the passive voice have been incredibly useful.
Unlike Google Docs, they do not offer to fix my problem for me. The onus for correction is placed upon the writer, something that promotes real learning in how to identify the passive voice, and experiential learning in how to fix it.
Since I have experienced a significant overlap between classes that ban AI and classes that explicitly do not grade from grammatical acuity, I believe services like the ones that scan for the passive voice could be of great help to students who are struggling to articulate their thoughts without the chance of AI directly or indirectly impacting their final letter grade. Students would learn and improve in targeted areas of deficiency without gaining an unfair advantage in areas that are actually graded.
When faced with reading long-form novels, AI has been equally valuable. While generative AI has the propensity to summarize entire texts and subvert reading as a whole, it can also generate tables of contents, provide contextualized definitions for words and antiquated idioms and suggest further reading to better understand the text. None of these actions allow a student to dodge the responsibility of reading and digesting a text, but alternatively provide an essential quality: approachability.
Turning back to the question of the assignment of long-form literature, the approachability of texts as well as the approachability of writing about texts, is paramount to the question of why students are unable to handle novels. Without proper preparation in high school, students are entering college without the necessary tools to sit down and read upwards of 200 pages while still maintaining a sharp eye for analysis and scrutiny. The solution for this isn’t just to force students to read novels anyway, nor is it to strip novels from syllabi either. Rather, professors should provide these students with the tools to approach these novels with a less intimidated disposition, and one of the most universally advantageous is generative AI.
The question, of course, is whether the floodgates of AI usage can successfully be just opened slightly — no more and no less. It is an unfortunate reality, however, that cheating and the circumventing of work have always been prevalent, a factor that must be weighed when considering the ultimate question: Can AI be effectively integrated into the teaching of literature? Thus, it appears that this latest hypothesis in the American experiment is just that: an experiment. At the end of the day, the variable in question is trust.
