How neural networks are changing the education system

Artificial intelligence (AI) is rapidly conquering the field of education. Computer vision, machine learning, neural networks, metaverse are now heard everywhere. Victoria Beresneva, director of the Autonomous Non-Profit Organization “Sports and Methodological Center “Department of e-Sports”, conducting comprehensive scientific and practical research in the field of high technologies, e-sports and AI in the interests of the largest players in the Russian video game industry, spoke about the depth and degree of penetration of AI into the Russian educational environment specifically for Invest-Foresight.

AI is a complex organized high-tech system that can emulate the intellectual abilities and creative activity of the human brain, the expert explains. Simply put, this is a machine that can act not only within the framework of predetermined algorithms and programs, but also approach tasks creatively. It can collect, summarize and analyze information, and sometimes even think outside the box more successfully than an ordinary human person.

One of the most striking indicators of the evolution of AI were its victories in logic games. When IBM-developed chess supercomputer Deep Blue won a 6-game match against world chess champion Garry Kasparov in May 1997, it was not yet AI in the literal sense of the word. It was just a super-powerful computer that won not through creativity, but thanks to the ability to store in memory and process lightning quickly information about millions of games played by people. Including chess fights with the participation of the grandmaster himself. But already in March 2016, a computer program based on AlphaGo AI, developed by the British company Google DeepMind, won 4 out of 5 games in the game “Go” against the Korean professional of 9th dan Lee Sedol. And it was already a truly self-learning AI, guided by nonlinear logic and capable of non-trivial moves.

Now AI has been able not only to learn itself, but to help people learn. In early February, Alexander Zhadan, an undergraduate student at the Moscow Russian State University of the Humanities, boasted that he not only wrote, but defended his thesis written using the ChatGPT AI-based neural network.

And already in April, a survey conducted by the Intelligent website found that of 1,000 Western high school, graduate and graduate teachers, 95% know about ChatGPT, 98% use it themselves, and 79% approve of students using the program. Moreover, 84% of teachers who approve of its use themselves taught their students to use the neural network. And 35% of educators believe that most or even all students complete tasks using ChatGPT.

The field of application of AI in education is expanding every day. Moreover, both in educational and administrative processes.

Using AI in online learning

The administration of learning processes can be greatly simplified and automated with AI. For example, AI can act as an assistant in adapting students, automatically sending them electronic forms and information. It can also assist in assessing students’ knowledge, tracking class attendance, reporting and auditing, and performing various administrative tasks. Products of Factspan Analytics can solve all these problems.

This also includes assistance to methodologists in the development of programs, lesson plans, as well as in the formation of educational content. It can be just a concept, idea or sketch, as well as fully developed text material, illustrations, audio and video files, presentations and much more. It is also worth noting the great benefit of language neural networks, which allow you to get acquainted with foreign experience as usual. A typical example of such a product is the same ChatGPT or Semantris from Google.

Training process support

AI offers increasing opportunities for learning and student development. Systems developed using AI are able to analyze data on student progress. Here, first of all, I would like to note the importance and need for the prompt implementation of virtual mentors and assistants. That is, digital assistants who can instantly give feedback to any request on the topic and suggest how to solve the problem without connecting a living person.

Adaptive learning has also become an important trend in education. This is when AI, based on student behavior and success, can offer them a simplified or complicated option for further learning. Thanks to this, the student can gain knowledge at the level where it will not be too easy and not too difficult for him at the same time and he will be able to complete the training. These actions also lead to the personalization of training, when the educational process adjusts not only to academic performance, but also to the personal preferences of the student.

Development of new forms of training

The use of AI in education leads to the creation of more and more interactive forms of learning. This process is closely related to the concept of “gamification”. That is, the introduction of game mechanics into the initially non-game – educational, professional or working -environment. Using AI, interactive interaction between game entities and players is organized. AI has a key role here. After all, without it, games would be less attractive to users. Gaming AI is based on computer graphics algorithms, control theory and robotics. It is essentially anything that the player does not control directly during a game session.

Benefits and possible challenges of using AI in training

A huge advantage of using AI in learning can be called adaptive learning and constant content updates. However, the current education system, unfortunately, cannot yet afford to use them fully. This is due not so much to a certain conservatism of the educational environment, but also to the banal lack of necessary technological infrastructure in some cases. It should also be borne in mind that due to round-the-clock support from digital assistants, a student can reduce the regulatory training time, saving his own and teaching time. At the same time, special emphasis should be placed on the economic component – reducing the burden on both the teaching and administrative staff. Moreover, this means optimizing the costs of the wage fund.

Eventual problems and challenges

However, along with considerable advantages of use of AI in training it has also essential shortcomings. Now the main problem of AI is that at generation of content it is always necessary to recheck correctness of the provided data and their compliance to the current legislation. Otherwise, comic or even tragicomic incidents are inevitable.

For example, not so long ago Elena Skarzhinskaya, the head of a chair of eSports of faculty of the game industry and eSports of the Sinergiya university, told how one of her students used the AI materials generated by ChatGPT in the final qualification work and inadvertently inserted the nonexistent cyber athlete there. And the research supervisor found it out.

In the USA the lawyer Stephen Schwartz on behalf of his client submitted  a claim against airline for the fact that his client during flight was allegedly strongly hurt by the metal cart for a delivery of onboard food. The lawyer provided to court the multipage report with legal precedents in favor of the claimant. It could become a strong argument against airline. However, any of affairs wasn’t revealed in databases of court as the report was made by means of ChatGPT. As a result, unlucky defender appeared before the court himself.

A few years earlier, Microsoft, in order to improve customer service, developed an AI-based chatbot named Tay for the Twitter social network. Initially, Tay was supposed to learn to speak like an average teenage girl and self-learn in the process of communicating with different people on the Internet. However, during the experiment, the situation got out of control. After some time, the neural network began to make statements of an antisocial and extremist nature. A similar misanthropy of AI caused a flurry of indignation among Internet users and almost let the developer company under a series of lawsuits.

Considering that in all of these cases the main fault was still on the person, one can make a not very pleasant conclusion. A lot of people are still not ready to engage directly with AI. And as the use of AI, including in education, expands, users are increasingly at risk of facing ethical problems.

Difficulties and even problems can also arise with the preservation of privacy, the spread of disinformation, including through deepfake technologies (an AI-based image or voice synthesis technique) and even in the field of social policy.

Already today, some organizations are beginning to reduce full-time employees actively, replacing them by AI. Even in the most unexpected areas.

“Today, artificial intelligence is already more effective than the capabilities of a specific specialist. Many professions are leaving human activity. For example, there are fewer church ringers, church carvers who make iconostases are replaced by numerically controlled machines. Artists are next in line. Perhaps the artistic level of the machine will be technically even higher than the artistic level of a person. The technologies that are now used in sketching projects for painting icons and churches were made manually 15 years ago, and now they ate made with the help of programs”, Dmitry Trofimov, head of the Tsargrad creative workshop, which recreated the paintings in the Cathedral of Christ the Savior in Moscow, complains.

In addition to painting and restoration work, he is also the general producer of the festival of contemporary church art “See and Hear”. One of the tasks of the festival, according to Dmitry Trofimov, is to “prepare for the offensive of the culture of artificial intelligence, which in 5-10 years will attack humanity. The confrontation for us will have an essential character, and the church worldview is the only thing that can resist the influx of this new culture.”

Even ChatGPT creator Sam Altman agrees with this formulation of the problem. Therefore, he calls on states to regulate the field of AI. Only training in new competencies can save from reduction. For example, the ability to interact professionally with AI to solve various kinds of problems.

A look into the future

The development and active introduction of AI into the education system, including online education, opens up wide opportunities to increase not only labor productivity, but also quality indicators of intellectual and creative work. As the number of educators who dare to experiment with this new technology grows, more and more innovative and non-standard ways of using it will inevitably appear. For example, already now AI interacts actively with the field of art.

At the end of last year, the world’s first exhibition opened in the United States at a branch of the New York gallery Bitforms, where the neural network co-authored artwork. The exhibition, entitled Artificial Imagination, presented works generated by the DALL*E neural network based on a text request. Moreover, they were created both according to detailed verbal descriptions of future paintings, and on the basis of extremely widely formulated requests. Among the participants in the exhibition were the creator of the Mighty browser Suhail Doshi and roboticist Alexander Reben.

The art exhibition “Art 2.0. Neuroscientist”, where works of art created by a neural network were exhibited, took place recently in Moscow. The exhibition featured works by Rembrandt, Leonardo da Vinci, Rene Magritte, Pablo Picasso, Salvador Dali, Casimir Malevich, Andy Warhall and other painters in interpretation of the neural network. As well as abstract works in futuristic and cyberpunk style, created entirely by AI. Therefore, it cannot be ruled out that AI could soon become an important element of learning in art schools and schools.

In general, if we talk about the use of AI in education, then it is just going through the approbation phase. Therefore, before starting to teach young people massively how to interact with digital technologies, domestic teachers must first study them carefully themselves.

Previous ArticleNext Article