Writing at university in the age of artificial intelligence


The arrival of artificial intelligence has put all educational models in check. The current paradigm of the teaching-learning process has been questioned and there are many voices that have supported the need to integrate everything that AI offers in the training of students, from basic education to higher education.

This is because AI has multiple functions that are useful in the classroom. It allows you to generate knowledge, review databases, design tools for the dissemination of knowledge, transfer content between multiple languages… But also automatically perform one of the most complex tasks that students face once they enter the university stage. : writing scientific texts.

The challenge of academic writing

Although writing is a skill that is worked on from an early age, writing an academic text requires knowledge that goes beyond that. Students have to become familiar with the modes of communication of a new discourse community and this implies identifying the linguistic-discursive characteristics of the genres that are produced in different disciplines.

Furthermore, writing is the means through which students’ knowledge is mainly measured (assignments, reports, reviews, written exams). As a consequence, it serves to measure success or failure in the acquisition of skills.

This challenge is not without difficulties and this has led to an increasingly widespread use of AI to respond to academic tasks. Faced with this, from the perspective of academic teachers, the consequences that this can have on the effective training of students are questioned.

Consequences of using AI for written tasks

One of the aspects that first comes to light is the one that has to do with academic integrity. If plagiarism was already the subject of debate, the difficulty in identifying who authored the texts multiplies the questions.

Despite the speed at which tools are generated, there are still no reliable digital resources that allow us to recognize which part of the text is authentic and which part is artificially generated. Therefore, using these resources would also have to involve teaching students that they have to be responsible for the products they generate .

In the same way, a problem arises that has to do with the very nature of academic writing, which has as its ultimate goal the genesis of new knowledge. To achieve this objective, it is necessary to select pre-existing reliable sources, contrast them and articulate a discourse based on them.

In this sense, the texts generated by AI are the result of predictive engineering that produces those words or ideas that are most probable. And probable is not synonymous with real. Furthermore, the quotes that appear do not have to be authentic, which detracts from the credibility of the discourse that is being created.

It is also important to keep in mind that when we provide data to the AI ​​to generate text, we are giving it away. All input and output from tools like chatGPT are stored.

The data may be reused with our full consent even for commercial uses. Security breaches in this sense are also questionable from an ethical perspective in the academic field.

The epistemic function of writing

Finally, there is an equally important dimension that affects not only academic writing, but the act of writing itself, and that is that it has an epistemic function . Writing is thinking, reasoning, developing, structuring and transforming knowledge.

A student who does not write does not think, does not communicate ideas, nor does he dialogue with his own thoughts. Let’s not forget that, when writing, multiple cognitive processes are activated that connect with short- and long-term memory. Therefore, by delegating the task of typing to a typewriter, students lose the ability to regulate their thinking.

In short, it seems unquestionable that students must be digitally competent citizens and that technology must be a tool to make their contributions to society more enriching. On the contrary, AI cannot help students have more resources to avoid the effort expected of them.

Author Bio: Mari Mar Boillos Pereira is an Adjunct Professor at the Faculty of Education of Bilbao at the University of the Basque Country / Euskal Herriko Unibertsitatea