August 5, 2020

Why your future teacher is not going to be a robot (but you will have to use one) | Economy



The question about how the future education will be is, in turn, a long succession of unknowns:the classrooms will disappear as we know them, converted into more flexible spaces? Will the screens banish the paper forever? And the teachers … will they end up being replaced by robots? Artificial intelligence has already reached education, accompanied as usual of apocalyptic affirmations. A little more than a year ago, British expert Anthony Sheldon ventured to predict that in 2027 intelligent machines would have replaced teachers. But the first essays point to a much less radical scenario, at least for the moment. No, robots are not going to replace teachers. But yes, the classes of the future will be a kind of tandem between humans and machines in which each of them will specialize in what they know best.

Sheldon's prediction for 2027 drew a disturbing picture. In his vision, the task of transmitting knowledge would fall entirely to robots, while teachers would be relegated to an assistant role (to maintain discipline in the classroom, help students, prepare the necessary material for classes … ). It does not seem like a feasible scenario in the short term. To begin, because the own artificial intelligence it is still far from being a true intelligence.

The experts distinguish between two types, the weak and the general or strong. The first is the one that already lives with us. It is about programming the machine to perform specific tasks, in limited ranges previously defined: recommend a song according to what you hear on Spotify, process your claims in the bank … or, in the case of education, resolve the question of when It is the date of an examination.

General artificial intelligence, on the other hand, is to provide the machine with a true intelligence that functions like the human, that is, capable of solving problems on its own and not in previously defined frames. Of having even emotions, personality or, for example, of master human language to be able to read the answers of an exam and decide if they deserve a 4 or a 10. It is a possibility still far. "It's the science fiction scene," says Guillem García Brustenga, director of trends at the eLearn Center of the Open University of Catalonia (UOC). "We talked about decades … although it is also true that we have been saying for 20 years that it is a matter of 20 years. That makes me think that nobody really knows. "

The future, therefore, is outlined in the form of a binomial between human and machine, in which the teacher will continue assuming the essential task of transmitting knowledge but can rely on the machine to download the most repetitive and anodyne tasks. There are already numerous examples of how artificial intelligence can be applied in educational centers to unblock, for example, the administrative and management part.

In the CEU Cardenal Herrera University, in Valencia, are about to release a new chatbot that is going to take charge of answering the most cumbersome doubts of the students: explain procedures, indicate tutoring schedules, clarify what paperwork must be presented … "These are questions in which the answer that the university or the professor can give has a little added value, but that the students need to solve with immediacy, "says Iñaki Bilbao, vice chancellor of International Relations. This intelligent assistant has involved more than a year and a half of development of the university along with Microsoft and Encamina, a technology consultancy.

The project is, however, a first step. The center wants to promote a combination of artificial intelligence and big data as a recipe to achieve that ideal that 21st century education pursues: personalized learning. Thus, they seek to collect all the possible data on how their students learn to develop predictive models of academic performance, capable of activating early warnings (for example, if a student misses class) that allow the center and teachers to implement personalized actions.

Artificial intelligence to adapt education to each student, but also to involve them more. "Technology enables active student participation in class. And that means that the teacher's attention is more exhaustive, "says Lorenzo Moreno, professor of University of La Laguna, who teaches a master's degree specialized in bringing technology closer to teachers.

The example of the Cardenal Herrera CEU serves to illustrate how the next step of artificial intelligence in education is to go beyond the scope of mere management to enter fully into the learning process itself. "In the administrative questions or even productivity there are more experiences by the analogy that can be established with the customer services of any company. But the uses with educational intentionality are much more complicated ", explains García Brustenga.

Algorithms to learn more

Here, the possibilities are enormous. The UOC eLearn Center has just published the research Chatbots in education, in which he recounts how intelligent assistants – just one example of tools based on artificial intelligence – can be used for educational purposes: to accompany the student, motivate him, help him practice specific skills (learn a language), simulate situations (a consultation with a patient), evaluate … The UOC has created Botter, a prototype of a robot designed to encourage its students with the study, capable even of showing disappointment if the student does not perform adequately. It is an idea of ​​the Department of Psychology, which wants to investigate if this type of stimuli serve to improve the motivation of the students.

Language learning it is fertile ground for this type of experiences. And the vocal recognition systems, already used for example in the smartphones, are your main tool. Although they also pose a challenge. The platform Lingokids, designed so that children can learn English, use algorithms to help define the content and activities offered to each child, depending on their level of English and their tastes. But it faces the difficulty that these speech recognition systems are intended for adults. "In children they do not work very well yet. With the added complexity that they are small children speaking a language that is not the mother tongue, "explains Carlos García Prim, its director of mobile engineering.

In ABA Español, a virtual language academy, turned to the prestigious Massachusetts Institute of Technology (MIT) and to the popular Alexa – Amazon's assistant in a small experiment: simulate a conversation between teacher and student to assess the level of English in just five minutes. The company already uses artificial intelligence, for example to assign real professors to each of its students. Your learning manager, María Perillo, ensures that far from replacing teachers, this technology makes them even more necessary. "The teacher now has a different role from what we used to think. Many times we just collect data and take the result of the machine for granted, but it is necessary to have the experts behind to analyze and interpret this data. Also to do what the machine can not: motivate the student, give him the feedback right … ", he lists. "It's not about using artificial intelligence for everything, but about using it in everything."

This division of tasks, together with the image of the teacher as a kind of assistant trainer, algorithms, data collection systems …, is a constant in the educational projects of artificial intelligence. Sometimes even in unsuspected lands. Of the more than 15,000 applicants that last weekend they faced the dreaded MIR test to get a resident doctor position, some had prepared for the exam with the help of an intelligent algorithm. After four years of development, the Editorial Medica Panamericana has recently launched Promir, a course on-line which is based on artificial intelligence to prepare the test.

Thus, the tool is able to analyze in detail each student to draw a personalized path: it tells you what to study each day, selects the questions and tasks according to what is most difficult and helps you review what you have already assimilated . That is the part where the machine is unbeatable. But it's not enough. "We apply artificial intelligence for what technology does well: to know the student, make a very detailed diagnosis of his profile and guide him in the process", explains Ignacio Ferro, its technological director. "But there are other tasks for which it is more effective for a professional to intervene: develop content, solve doubts, prepare the student psychologically. For that there are specialized tutors. "

But as in any scenario that moves between an understandable present and an almost science fiction future, ethical questions arise. The UOC report highlights one question: what is the final goal of the machine? If it is that the student learns more, the researchers point out, the risk is that artificial intelligence poses too difficult challenges that lead to suspense. If the purpose is to approve, you can set standards too easy so that the student finally does not learn. And if the goal is to increase enrollments, the game already enters the debate about whether technology is a means to improve learning or an end in itself to, for example, use as a tool marketing. "We will have to reach a commitment, even at a social level, about what we want from all this," summarizes Guillem García Brustenga.

.



Source link