I’m a college writing professor. How students should use AI this fall

People reach all kinds of metaphors to describe their relationship with AI. For some, AI is like a mainly reliable trainee. For others, it is a virtual assistant. Increasingly, chatbots like Chatgpt move in the role of companion, therapist and even a romantic partner. As a professor of university writing, I came to consider AI as a collaborator: an archive of knowledge that returns. But as an alcoholic sober myself, I can’t help but imagine it as a high-level drunk: it can sometimes seem brilliant even when he has no idea what he is talking about.

I can tell you stories about the ways that IA went through when I needed it, making myself save hours by doing banal tasks, rereading my writing or conversing on my latest obsessions of research. But there are then these other moments when he resides in a joyful tone, when he seems to not understand a word that I say, but continues to speak rather than admit that it is bad or that he has no answer. As a few weeks ago, when I asked Chatgpt to transform my written remarks for an academic conference into a game of slides. My speech concerned literary journalism, and he proudly offered me a presentation on luxury trips to Brazil.

Excluding incidents like that give me many stories to share with my students. But even if I think AI undermines some of the most important human reasons to writeNot all kinds of writing are the same. To write, we often have to look for first, and after writing a project, we need critical comments. Instead of adopting a reactionary AI approach, I want to explore with my students how it can be a useful collaborator in this process.

Discuss with archives

A large part of university writing is based on research and reading, a process that forms the mind to organize information and to think logically. But using new technologies for this process does not mean that we do not yet do critical mental work. Just in my life, these technologies have radically changed: we have gone from library cards and microfiche catalogs to online databases like Jstor and Google Scholar. These tools do not require less reflection – they simply accelerate part of the brainstorming and collecting information, and they widen the quantity of knowledge that we can consider.

Because I had witnessed this rapid digitization of research and writing tools even before AI, I am more inclined to imagine that AI can be a collaborative research partner. In my field, for example, literary researchers spend hours painting thanks to primary sources in libraries and archives. Digitization has already facilitated access and AI can make them easier to analyze.

Lately, I realized that we could think of talking to an AI chatbot who does not like to travel through an archive, but as conversation with one. Before diving into a more intensive work, we can have a conversation focused on research with a “spirit” which has at least a general idea of what exists. A few weeks ago, I used my limited access to the advanced vocal function of Chatgpt to ask me if he thought that this idea of discussing with the archives was a reasonable way to understand what is happening when I converse with AI. He replied: “When you talk to an AI like me, you access a large amount of information and models derived from human knowledge to a certain point.” He also covered a little: “It is important to remember that even if I can provide information and ideas based on this knowledge, I do not have human experiences or conscience. So, even if it can want to converse with a vast reservoir of knowledge, it is always good to also consider human perspective and the context.”

However, over our conversation and my questions have become more sharp, I could ask him to provide references and places where I could go for more in -depth reading. Since this first provisional conversation, this conversation before writing with AI is part of my workflow. I have always found it easier to find my ideas through dialogue, but few people are interested in hearing my hay ideas. This is why I found that speaking through ideas is one of the best uses of AI for writers.

Create your own mini-archive

Although the discussion with AI has proven to be useful for the generation of ideas – and the fact that it retains a transcript facilitates the reference later – there are an increasing number of tools based on AI designed to help with the more intensive research phases. At the end of the fall semester last year, a student sent me an email asking me if I had heard of Google Notebook. I did not do it, but when I opened the link, I obtained the concept almost immediately. NoteBooklm takes the idea of talking to the archives at the next level: the archive with which you discuss is the one you assemble with sources for a particular project, which AI can also help you collect to start.

Preparing for my recent conferences, I threw 25 pdf that I had assembled and stored ZoteroMy favorite quotation manager, in the notebooklm Note interface. He quickly “read” them and provided a summary that started: “These sources discuss the philosophy of ordinary language, focusing mainly on the work of Wittgenstein, Austin and Cavell, and his relationship with other philosophical and literary movements such as pragmatism, transcendentalism and deconstruction.” Under the summary is a field of entry of text which encourages me to “start typing …” and provides suggested invites like: “How does the philosophy of ordinary language challenge the traditional philosophical approaches of meaning?”

On the right side of the page, in an area designated “studio”, I am invited to create an audio preview, which takes the form of a podcast, with two voices – a male, a woman – pleasant on my chosen subject. If I use interactive mode, I am treated as an appellant in an old end of evening radio program. I receive compliments for my excellent questions and answers according to the documents I provided. The part of the podcast is not yet great; It is frightening, but I can imagine that it improves and becomes more useful. Notebooklm has other useful features: it can create a “mental card”, a study guide, a briefing document, a FAQ and a chronology. I will continue to use it and suggest that students do it too.

Meet the creators of the industry, the contributors and the opinion leaders who have teamed up with the award -winning Mashable staff to publish comments in the first person of the lived experience.

A writing tutor not quite

Another way in which AI could be useful to students ‘writers is its ability to provide instant comments on students’ writing. When I asked Chatgpt about this concept, it encouraged me to “think of AI as a writing tutor available 24/7” with the warning that he “misses the personal touch and nuanced understanding of individual students that a human tutor provides”.

I stuck in the full text of one of my Previous mixture stories And asked for suggestions. He seemed to know what we refer to in the workshop sessions of homologists in class as the “sandwich with compliments”: criticism fell between two compliments.

He said to me: “It is a convincing and eloquently written play … Your voice is authentic and thoughtful”, before offering “some suggestions to raise the play more”. Again, he started with “the forces to keep”, followed by “suggestions for improvement”, in particular “tightening openness”, “strengthening transitions” and “considering a stronger conclusion”. He also had some “minor style changes” to suggest. Finally, he provided a global note: 9/10.

It may be all the compliments, but I became greedy. I stuck in another test (9.5/10), then the conference on which I worked. The global impression has started well: “Your article presents a convincing argument for the value of literary journalism which focuses on” ordinary “and” daily “.” It’s true, although I have never used the word “daily”. But then – I should have expected – it came out of the rails. “References to fundamental figures (for example, Bateson, Becker, Carey, Geertz, Tuchman) and contemporary examples (for example, Kiese Laymon, Eliza Griswold, E. Tammy Kim) helps to situate your argument in a well -informed learned setting.” I do not refer to any of these figures as fundamental or other.

I drew his attention on this subject, and he said that I was “absolutely right” and thanked me for reporting it. His explanation, however, was always confusing: “I am wrongly based on my answer on hypotheses or ideas cache other academic discussions on literary journalism, not your specific article.” I study literary journalism; The names that Chatgpt has abandoned belong to writers, but they are not scholars in my field. However, after having corrected it, we came back on the right track and that provided comments, using the compliment’s sandwich again.

I do not know what to do with the fact that Chatgpt has resisted myself much better with my more journalistic writing as opposed to the academic, except that it provides me with another opportunity to urge prudence when you help students think about appropriate AI uses to complete – rather than replace – the writing process.

In the end, I like the notion of AI as a communication, although something that takes place occasionally, leading it to a too flat and simple lie. I am absolutely for the idea that by discussing with a chatbot, writers can approximate something like talking to a whole series of human knowledge, in particular with a tool like notebooklm which allows writers to “teach” AI on a subject before discussing.

AI as a collaborator calls me, even if I have to approach it with a skeptical skeptic, always prepared for the next time that it will drop me.