The Moment of AI truth:the results of the IOM 2025 Olympiad

The International Youth Olympiad (IOM-2025), organized by the Higher School of Economics and providing young foreigners with the opportunity to enroll at  a Russian university, for the first time offered future journalists and media industry workers to use generative artificial intelligence (AI) to complete competitive tasks. The experiment showed that participants are coping well with new technologies, but also revealed new challenges in the usage of AI in the field of media.

© Olga Zotova

The International Youth Olympiad of the Higher School of Economics is an opportunity for foreign applicants to get a budget place or grant of up to 75% of the university tuition fee. This year, the Olympiad, which is being held for the tenth year, attracted a record number of participants – more than a thousand people applied for the tracks  «Journalism» and «Media Communications » that is twice as much as last year (500 registrations). Residents of any country who do not have Russian citizenship could take part in the competition, starting from the pre-graduation 10th grade (or from the 11th, if the school programme included 12 grades).

Advantages and goals of the new format

This year there have been significant changes in the format of the Olympiad. For the first time, on two tracks «Journalism» and «Media Communications» tasks included the mandatory use of AI. The purpose of this experiment is to evaluate the practical mastery of advanced technologies that will be necessary to create high-quality and competitive media products in the future.

Ivan Knyazev, Chairman of the Methodological Commission for Journalism, explained that the Higher School of Economics is actively developing the field of artificial intelligence and integrating it into the educational process. According to him, the Olympiad was an experiment in the usage of generative models in journalism.

Ivan Knyazev
Chairman of the Methodological Commission for «Journalism»

«An important goal of IOM-2025 was to determine the level of familiarity of future journalists with artificial intelligence in order to adjust training programmes. It is important to understand which courses on working with AI in journalism are necessary — both introductory and specialized (for example, on the stylistics of texts created using AI).»

Ivan Knyazev

Alexander Kupriyanov, Chairman of the methodological commission for Media Communications, noted that the idea of mandatory use of AI in IOM-2025 tasks is due to objective reality.

Alexander Kupriyanov
Chairman of the Methodological Commission for «Media Communications»

«Participants have free access to the Internet and, consequently, to AI, which makes it impossible to ban the usage of it. Instead of banning it, we, the methodologists and organizers, decided to make AI an integral part of tasks, turning it into a tool for solving Olympiad tasks.»

Alexander Kupriyanov

The Role of AI in tasks

In the first round, participants of the «Journalism» track were asked to write an extended news article based on the provided text (press release, article, interview). Participants could use various AI models (ChatGPT, YandexGPT, GigaChat, Perplexity AI, etc.), choose a tool for processing information and creating news independently. In addition to the note itself, it was necessary to show the used prompts.

According to the Chairmen of the methodological commissions, the tasks were formulated in such a way that it was impossible to complete them without demonstrating contestants' own knowledge and skills. To get the maximum points, participants of the Media Communications track had to write a podcast script, showing a good knowledge of the topic, the basics of drama and stylistics, the ability to compose high-quality requests (prompts), as well as a richness of language. The news article evaluated the accuracy, texture and completeness of the topic, compliance with genre requirements, knowledge of AI methods, as well as stylistics and literacy of the Russian language.

Alexander Kupriyanov explained that participants who knew the basics of drama well formulated requests to AI much more accurately than those who did not have this knowledge:

«They formulated requests not just as «add drama», but more precisely, for example, «create a cliffhanger», «formulate a narrative with a beginning, development, climax and denouement». By the quality of prompts, we judged the participants’ depth of knowledge. Many were limited to common phrases like «generate an 8000-character podcast script», while the most successful contestants demonstrated an understanding of the prompt structure using frameworks that took into account the model's immersion in the role, context, and target audience. Some have used AI to refine already generated text, adding emotional swings or more relevant examples. However, complete dependence on AI did not work here: even the most powerful models, such as the OpenAI chatbot, are limited in the relevance of information».

A similar situation was observed with the task of writing a news article. Participants who knew the news structure (headline, lead, quotes, main text, background) formulated requests that considered these components. For instance, they required AI to start the lead with a subject and predicate. Thus, the assessment of participants' knowledge was based not so much on the quality of the AI-generated text, but on their ability to formulate effective requests to the neural network. This has become a key difference between IYO-2025 and traditional Olympiads.

Ivan Knyazev emphasized: AI is just a tool. Human involvement was needed to set the tasks and check results properly: «For instance, in the second task, participants were required to define what «analytics in journalism» is. Participants who did not understand the specifics of this term provided definitions from other fields, indicating a lack of knowledge. If the term was understood correctly, the AI would help them formulate an accurate definition». 

What did the Olympiad show?

The results of the Olympiad revealed a wide range of competencies: from works that impressed the jury with their quality and made them think about the future of the profession to works that highlighted typical mistakes: inability to work with stylistics, to highlight the main points from large amount of information. According to what Knyazev notes, many participants demonstrated an average level of competence, indicating the need to further develop practical skills in interacting with AI in journalism:

«The experience gained will become the basis for further studies and improvements of the Olympiad’s tasks. We are planning to continue experimenting with the usage of AI in future specialized contests».

Participants could edit the responses generated by the neural network, although some did not take this opportunity, limiting themselves to simple text generation without further work on it.

«We expected that contestants would ask AI for help in structuring information and building logic, but there were few such cases»,  noted Alexander Kupriyanov.

The contestants who successfully completed the tasks of the first round faced the second stage — an oral interview. This allowed an assessment of general cultural level, knowledge of the media environment, and the ability to argue their position.

Ivan Knyazev, Chairman of the methodological commission for the «Journalism» track, Academic supervisor of the «Journalism» programme

«The general outlook of many participants turned out to be extremely limited. Their knowledge was confined to the school curriculum, which, it seems to me, is due to the enormous pressure of preparing for final exams, which almost completely absorbs the time of high school students. Some participants struggled to answer basic questions in history and geography, raising serious concerns about the quality of modern school education. For example, the question «Which country doesn’t exist: Sweden or Switzerland?» confused some. Journalism still requires a broad outlook and universal knowledge».

According to the results of an anonymous survey, participants of the Olympiad highly appreciated the potential of AI, noting its positive impact on the quality of their work. However, contestants also pointed out difficulties in formulating requests. Looking ahead, they expressed their wishes for a detailed analysis of methodological materials that include not only explanations but also practical examples of using AI to successfully complete tasks. Participants were particularly interested in samples of successful works and recommendations for formulating requests. They would be helpful for understanding how many prompts and in what form should be included in the assignment.

IYO-2025 has demonstrated a growing interest among foreign youth in media and innovative technologies. The Olympiad not only identified talents but also emphasized the need to adapt educational programmes to the realities of the modern media environment, where AI is becoming an important tool.

Author: Azaliya Mukhametshina, a third-year student of Bachelor's programme

«Journalism» 

Translation: Ekaterina Erokhova, a second-year student of Bachelor's programme

«Media Communications»