ChatGPT and education: a new enemy or ally for teachers?
How to fight against an apparently undetectable application that is capable of generating, in less than a minute, a 600-word essay on nuclear energy, the migratory crises of the 20th century or gender inequality, and also do it with language completely natural? The irruption of generative artificial intelligence in the educational field is taking its first steps and has already meant a real turnaround around the world. When, at the end of 2022, OpenAI launched the popular ChatGPT, an artificial intelligence system capable of generating original texts, answering questions and solving a multitude of tasks, it took them five days to exceed a million users, and 100 million in just two. months. Faced with the danger of a tool that facilitates plagiarism and cheating, the alarms in schools and universities were not long in sounding, to the point that there are already many cases (the State of New York and the schools of Seattle, in the USA , and Australian universities are just a few examples) where there was a quick move to ban its use.
The question, however, is not whether this phenomenon can be contained, but rather how to incorporate its use productively: according to a recent Impact Research survey, one-third of US students ages 12 to 17, and the 51% of teachers already use this tool (a percentage that reaches 69% in the case of Latino or black teachers), either to plan their classes, find creative activities or develop a base knowledge with which to start their studies. lessons. Their conclusions could not be better: 89% of teachers and 79% of students who have used ChatGPT think that the impact has been positive. The wave is unstoppable: Microsoft has already announced the addition of its own generative AI, Copilot, to Office 365; and Standford has released Alpaca, an AI based on the open source ChatGPT. “Neither students, nor teachers nor journalists will start something with a blank page again; they will ask ChatGPT for it”, ventures Guillem García Brustenga, an expert at the UOC’s eLearning Innovation Center.
fight against plagiarism
For García, the change in methodology is mandatory, and education must coexist with artificial intelligence (AI) in the same way that it coexists with Google or Wikipedia. “The problem in the short term, of course, lies in what we are going to do to prevent the student from copying. That’s clear. But we are not going to stay here, rather we are going to take advantage of this opportunity to make changes that should have already been made”, adds García. “Let’s see: if a teacher asks what were the causes of World War I, and that is the exam question, this could already be copied from Wikipedia… Maybe what is wrong is the question. What happens is that now it is easier, because we only have to go to ChatGPT and ask him, because the answer will be original and will not be detected by the anti-plagiarism systems”. The solution, he points out, will necessarily involve reinventing the role of the teacher and adopting new teaching and evaluation methodologies, so that they serve not only to avoid cheating, but also to learn more.
Instead of handing in a copied or memorized text from Wikipedia, says García, the teacher can ask them to make an infographic, to contextualize it with their reality or to use portfolios to document their learning process, from the moment they don’t know anything until they know a lot. “Paradoxically, the incorporation of AI through tools such as ChatGPT is going to force us to reflect in depth on what we want to do, how we want to do it, how we are going to evaluate it and what kind of skills we want to cultivate”, he maintains. Joaquín Rodríguez, director of Design, Innovation and Technologies for Learning at the SEK Educational Institution. And it is that, instead of trying to prohibit its use, he suggests incorporating it into teaching as a starting point for discussions and debates in the classroom: “I would ask the students to use it and ask questions. And all that information that the AI gives us would later be put on the table in the classroom, to be subjected to scrutiny, debate and criticism by the students. That they reflect on it and learn to synthesize its content, to deepen and contrast the different answers they obtain… They must be sure of their successes, their limitations and the biases they may have”.
That is precisely one of the biggest dangers of ChatGPT: the assumption of veracity, because it will never admit that it is not telling the truth about something. “One of ChatGPT’s mistakes is that it has what are known as “hallucinations”: it always claims to be true, so if it doesn’t know something, it makes it up,” warns Rodríguez. For this reason, from the pedagogical point of view, he recommends preserving and reinforcing fundamental cognitive skills, such as the ability to verify and evaluate any statement that artificial intelligence gives you; keep in mind that the final responsibility for the opinions of an AI remains human (AI can be used in a medical consultation, but the responsibility of the diagnosis will always rest with the doctor); and make critical, productive and ethical use of technologies.
Reinventing the role of the teacher
The turnaround that the advent of artificial intelligence has entailed is a challenge before which educational institutions cannot, and should not, look the other way. “A more or less uncritical assumption of technology, that is, letting everyone do what they want, would lead to uncontrolled chaos. For this reason, it is necessary to accept it meditatively, which is the position that we, for the moment, assume ”, explains Rodríguez. And that necessarily happens by facilitating a radical change in the role that the teacher has in the classroom. “The role of the teacher should have changed a long time ago. When I studied, the teacher was the one who had the knowledge; I would go to class, he (or she) would tell me things and then say: “Now repeat it”. But that’s not the case anymore: when knowledge is everywhere, it doesn’t make sense for me to go to class just to listen to someone who reads to me what I can look up on my own”, affirms García.
For the UOC expert, the teacher’s role must be that of “someone who helps me interpret this information and learn with it. In the world of ChatGPT, it will be the same, and this teacher who was explaining something that he himself had read in a book, will have to change”. And the same happens with the evaluation: instead of asking the students to repeat what they have been told, it will be necessary to ask in another way, “to do collaborative work or even oral tests that can be done synchronously or asynchronously, in which the students record themselves answering a question, reasoning or explaining what they have asked in the chat”, he adds.
While companies like TurnItIn, specialized precisely in detecting plagiarism in academic papers, claim to be already working on the development of a tool capable of detecting the use of generative AI, experts are skeptical: “At best, it could analyze some ChatGPT and claim, with 80% probability, that it is an AI. But I can’t suspend anyone for this. Also, if you take the text from the AI, change it and make it your own, that can no longer be detected by anyone. It is a complex issue”, says García.
Challenges in the world of work
If we have been hearing something for a long time, it is that artificial intelligence would come to replace the human being in those more mechanized tasks and trades. However, generative AI applications such as ChatGPT or Copilot show that they will not remain alien to the more creative professions either. “They are capable of developing (and executing) code themselves, and of developing entire texts and novels. There are already publishers based on artificial intelligence, or books that are being uploaded to Amazon; and if you ask them to make you a painting in the style of Picasso or Monet, they will do it too. For this reason, one of the most important questions being asked now is what skills will need to be developed”, says Rodríguez.
Now, will the AI be able to take jobs away from us? For García Brustenga, it will not be exactly like that: “I believe that it is not artificial intelligence that will take away those positions; Rather, it will be the professionals who know how to use artificial intelligence who will take jobs from those who do not know how to use it. That is the real issue. For this reason, we have to train future professionals while retraining ourselves, to know how to use all these tools; and even then, we will suffer from the speed of change. It will not be technology alone that will take our jobs away.”
TRAINING THE COUNTRY in Twitter and Facebook
subscribe to newsletter of Training of EL PAÍS
Five Days agenda
The most important economic appointments of the day, with the keys and the context to understand their scope.
RECEIVE IT IN YOUR MAIL