The text by Prof. Jan Fazlagic, a professor at the Poznan University of Economics, is an excerpt from a publication dedicated to the impact of artificial intelligence on education in Poland. From the IQS’s perspective, the development of AI is also an opportunity to introduce AI qualifications into the system, which teachers in Poland will need in the future.
The development of artificial intelligence as a challenge for the education system
By design, artificial intelligence (AI) is supposed to help humans understand the world, reason, plan, communicate and perceive. Thus, it is supposed to be an enhancement of human intelligence and replace humans where their cognitive abilities are insufficient to achieve expected goals. However, when discussing artificial intelligence and its applications in the education system, we should take into account the fact that the education system relies heavily on shaping human intelligence. Thus, a kind of dilemma arises: can and should artificial intelligence intervene in the process of shaping human intelligence? What might be the long-term, deferred effects of such a process? What might society look like 50 years from now, in which human intelligence (whatever that will mean 50 years from now) will be a sort of product produced by the artificial intelligence used in the education system for teaching? Will the student “unburdened” by artificial intelligence actually benefit from such a solution? Or, on the contrary, is it likely to be counterproductive to apply to the education system the philosophy of using artificial intelligence in business, which is based on relieving humans of burdens? If artificial intelligence is supposed to facilitate business-customer interactions, is it desirable to apply the same approach to the education system-student interaction? The philosophy behind providers of AI solutions emerging for business needs may prove useless or even harmful if unreflectively applied to the education system. That which is “optimization” and “efficiency enhancement” in business does not necessarily mean the same for the learning process, e.g., teaching poetry or instilling a love of art will not be more effective if the learner learns about twice as many artistic works in a unit of time.
The challenges of using educational theory in developing artificial intelligence algorithms
The use of artificial intelligence in education is probably fraught with a greater set of risks than in other areas. If artificial intelligence wrongly advises a buyer to purchase a piece of furniture from an online store, the consequences of such a purchase will be relatively harmless compared with the introduction of an ill-designed teaching algorithm in school. The risks of developing inappropriate algorithms can be divided into the following categories:
- The risks associated with the use of the wrong or incorrect theory;
- The risks associated with negative side effects or unintended consequences of the AI-driven learning process;
- The risks associated with the incorrect inference by the algorithm and application of inappropriate tasks for the student.
How can AI harm young people’s education?
For example, an educational program using AI Kidsense (Irvine, California) allows a child’s speech to be transcribed into text for note taking. The challenge here is the difficulty of understanding the speech of young children. So the software is able to understand a toddler’s unclear speech. But the question arises, how will such a solution affect the child’s development? We do not have the results of a study on this
subject, but we can hypothesize. In the days when there was no such software, the child developed its ability to communicate by receiving feedback from the environment. By sensing which words and utterances were understandable to the environment and which were not, it spontaneously tried to be more understandable to caregivers, to improve its pronunciation, to adapt to the expectations of the audience. [...] Therefore, won’t Kidsense programming cause harm to the child’s development? In the extreme case, we can imagine children who will grow up to be babbling adults, or at least adults with serious speech defects.
Another example showing the potential risks of using AI in education is illustrated by the QUIZLET films of the Quizlet Learn educational program (San Francisco, California). They assist learners in synthesizing the material they are studying (literally helps take the guessing out of what to study). This platform, using machine learning and data from millions of lessons (study sessions), points learners to the most important material. Again, this raises the question: To what extent does broadly relieving and helping students really serve to develop their cognitive competence? Case in point: what will the world look like in which adults will not be able to find the right information in the surrounding world on their own, because this task was done for them by artificial intelligence during their formal education? Such an approach is a simple way to stifle the development of creativity in a student. [...]
Overview of AI applications in education
There is an ongoing debate about whether artificial intelligence will replace the teacher. On the other hand, the number of potential applications of artificial intelligence in the education system is much broader than just “replacing the teacher”. AI can act as both an autonomous teacher and a human-teacher assistant in the teaching process. We should not limit the field of applying artificial intelligence in the education system to activities oriented only directly to the student. On the contrary, artificial intelligence is not only able to replace the teacher, introduce mechanized elements into the learning process, but it can also help manage the student’s educational back-office. [...]
AI can help teachers detect learning problems in students, adapt material to individual student needs, or simply be used for grading. AI can also help manage lessons and the education system, for example, by predicting certain trends, student behavior and phenomena in advance. AI can help analyze student progress in real time and make recommendations on how to behave toward specific students. [...] Among the applications, several subcategories are distinguished:
- Artificial intelligence to support the learning process,
- Artificial intelligence to support the teacher in administrative processes,
- Artificial intelligence to support the management of school education at its various levels.
This systematization is important because it helps guide further research work and artificial intelligence development strategies at the level of the solution provider. Each of the above-mentioned applications has the potential to improve the situation of students, but from a different perspective.
AI should help, not harm, teachers
Because artificial intelligence exhibits many human characteristics, such as the ability to learn, think critically and solve problems, it raises many emotions. One of them is the fear that AI will replace teachers in the future. However, from today’s perspective, AI should primarily focus on repetitive tasks. With it, teachers would be able to spend more time individualizing their approach to students, for example. A Polish teacher will not have to correct tests, because artificial intelligence will give feedback to the student more efficiently. Within the education system, it is also important to consider whether the retreat from standardizing the learning process that the potential use of artificial intelligence could provide has negative side effects. Schools and the education system are places where socialization occurs, the interactions between students allow them to get to know each other, create a common value system, sense of belonging. Culture is created in schools, something that binds society together. Culture is neither “right” nor “wrong” – unlike mathematics, artificial intelligence should therefore primarily focus on the formation of knowledge and skills that have an objective reference point.
The article is a part of a broader publication on AI. You can download the entire publication using the link below (the text begins on page 24).
Author: Prof. Jan Fazlagić – Professor at the University of Economics in Poznań. One of the pioneers of knowledge management in Poland. Fulbright scholar and Marie Curie Research Fellow. Participant and manager of many national and international projects, including at the Polish Agency for Enterprise Development, the National Institute of Local Government and Climate-KIC. From 2011 to 2015, responsible for development and innovation at the Vistula University Group in Warsaw. Author of more than twenty books and several hundred scientific and popular science articles. He is interested in the possibilities of using artificial intelligence in the education system, teaching and in the field of (artificial) creativity. Speaker at numerous conferences in the field of education (including climate) and local government.