Why did the AI ​​come so late?

Artificial intelligence in education: opportunity or risk?

When machines teach

Digitization is making profound cuts in every area of ​​life.
Much is changing at breakneck speed. These upheavals should accelerate again when applications based on Artificial Intelligence (AI) enter our lives. AI will also have an impact on education

As with everything new, including digitization, fears arise that are often unfounded. Digitization and AI certainly threaten numerous workplaces whose activities can be automated. However, there can be no question of abolishing the teaching profession, although it will continue to change. Digitalization has made learning content more dynamic and multimedia. Artificial intelligence, however, will advance into completely different dimensions. The kettledrum should become an inspiring experience. However, the robot pedagogue clearly belongs to the realm of fiction.
Instead, various projects are researching to what extent the education sector can benefit from AI and how it can support teachers. During the coronavirus pandemic, many teachers experienced first-hand how complicated and time-consuming it is to look after students in homeschooling. In the future, they could receive support from Natural Language Processing (NLP). Techniques and methods for the machine processing of natural language are hidden behind NLP. AI systems with NLP are able to understand freely formulated answers. You can rate them and, on this basis, give feedback to those who have solved a word problem online, for example. This allows you to query what you have learned or to determine your understanding of text or a foreign language.

Language and behavior

NLP is used, for example, in the EU-funded L2Tor project. As part of the project, in which Bielefeld University is also involved, European scientists are developing a robot that will help preschoolers learn a second language. The beginnings were quite laborious, because on the one hand language support in preschool age is not standardized and on the other hand the AI ​​speech recognition is based on the vocabulary of adults. The robot first had to be trained.
The researchers found, however, that the learning success of the children improved compared to conventional learning of a second language, the better the language and behavior of the robot were adapted to their needs. The use of the robot was particularly helpful when learning vocabulary. In particular, the children were better able to memorize vocabulary that the robot made explanatory gestures. For the educators who were allowed to test it as part of the research project, it represented a great deal of support, if only because there is often simply not enough time for individual language training in preschool age.

The intelligent textbook

The time saved or the time gained in order to be able to take better care of the students is one of the arguments used by AI advocates. Instead of just checking the result of a math problem for correctness due to lack of time, the AI ​​should analyze why the result is wrong, for example whether the mathematical formula to be used was not understood. This goes beyond pure knowledge inquiries with NLP and ultimately means individual support for each student. For this purpose, the German Research Center for Artificial Intelligence (DFKI) is developing an anticipatory textbook for computers and tablets in cooperation with the Technical University of Kaiserslautern in the HyperMind project.
The contents of the textbook are based on the skills and needs of the learners. The researchers use an eye-tracking technique for the textbook, with which the reading directions can be recognized. This makes it possible to determine where the gaze lingers longer or which content is read repeatedly. The researchers want to analyze this data in order to be able to draw conclusions about learning behavior and learning progress, for example. "In the future, technology can help to identify at an early stage whether a student needs support on a topic, for example because he has not yet understood it correctly," explains computer science professor Dr. Andreas Dengel, who together with physics professor Dr. Jochen Kuhn heads the HyperMind project.
In addition, with the help of the intelligent textbook, one can see where the learners have particular interests. "If, for example, he looks at a certain word more often in the sequence, the system could provide him with further information, for example via the Internet browser," adds Kuhn. In this way, the static book structure is transformed into associatively linked knowledge modules.

Support vs. surveillance

The intelligent textbook demonstrates one of the strengths of artificial intelligence: learning is individualized. With the help of AI, learning profiles can be created that contain the respective strengths and weaknesses of the students, so that the AI ​​can set individual tasks based on the learning needs. The basic technology behind it is based on intelligent face recognition.
However, critical voices warn that it is only a small step from support to surveillance. The best example: China. At Elf High School in Hangzhou, facial recognition is used in classrooms to check whether the students can follow the lesson. Those who seem overwhelmed receive advice on how to increase learning success. Those who show above-average performance receive tasks that correspond to the respective level.
But that's not all: you can also borrow books or distribute food using facial recognition. Parents receive a nutrition report - and suggestions for better nutrition for their children. Those who arrive too late can no longer simply sneak into the school premises, because the gate only opens after the start of the lesson after successful facial recognition. And the teachers receive a message on their smartphone who asks for admission in front of the gate. But that went too far even for the Chinese. There were protests among the parents. However, it can be assumed that the Chinese government will continue to push the use of artificial intelligence in schools.

Don't ignore AI

In view of the potential of AI-based applications, it is certainly right that Germany takes a critical stance towards digitization - this can be seen in the discussions about data protection, which do not exist on this scale in other countries. Correspondingly, people in this country are also critical of artificial intelligence. But: it will come. The unanimous opinion of science is that ignoring AI would be fatal. The Tübingen Competence Center for Machine Learning, which launched the Federal Artificial Intelligence Competition (BWKI) under the patronage of the Carl Zeiss Foundation in 2018, sees it that way. In the past year, over 2,000 schoolchildren, teachers and interested parties took advantage of the BWKI's offers.
Schoolchildren aged 13 and over are eligible to participate. In a free online course, you will gain insights into the basics of AI. With a weekly programming task you can deepen your own programming knowledge. Participation in the online course and the implementation of the programming task is voluntary. But: The school with the greatest participation in the AI ​​course is chosen as the AI ​​school of the year. A bionics educational kit attracts as a prize at the final.

The intelligent chicken coop

Last year, over 150 school students submitted their machine learning project ideas to the competition. The BWKI finalists have developed, among other things, AI systems that recognize fake news, determine leaf diseases on grapevines or translate spoken words into a written text by lip reading for deaf people. In addition, prototypes such as an intelligent white cane or a smart feeder for the chicken coop were developed.
“With the federal KI competition, one of the most important topics of our time is coming to schools,” explains Theresia Bauer, State Minister for Science, Research and Art in Baden-Württemberg and, as chairwoman of the Carl Zeiss Foundation Administration, the patron of the BWKI. "It brings teenagers and young adults into contact with artificial intelligence at an early age - not just as passive consumers, but as active designers."
Prof. Bernhard Schölkopf, head of the Max Planck Institute for Intelligent Systems and one of the initiators of the BWKI, can only agree with this: "Our future is shaped by our children, so our support has to start with the children." In promoting AI, it is therefore important that children and young people understand how AI systems work so that they later know what they are dealing with - and not believe that Alexa is a friendly woman.

additional Information

bw-ki.de | www.dfki.de


Source: BildungsSPEZIAL 1-2021