#UNDPDigitalTalks: How We Became Familiar with AI

The World Ahead and AI in Education

February 9, 2024

In December 2023, United Nations Development Programme (UNDP), Petlja Foundation and the Ministry of Education joined forces to organize the conference “World Ahead”, so that, together with experts in the field of artificial intelligence, they could help a couple of hundreds of teachers from all over Serbia to get to know generative artificial intelligence better. 

We asked multiple questions – how to use AI in classrooms, what is the attitude among teachers, and what among students when it comes to using ChatGTP, and whether we would be replaced by robots, but we also addressed the ways in which artificial intelligence may help us become more competent. 

As a wandering reporter at the conference on artificial intelligence in education, I had a first-row seat to observe the emergence of an interesting phenomenon that was involving all the participants, throughout the day.

The phenomenon in question was the rubber band effect.

Let me clarify. We need to go back one year to a lecture given by two AI scientists, Tristan Harris and Aza Raskin. 

At the time – unlike today – few people truly dealt with the phenomenon of generative artificial intelligence embodied in ChatGTP, DALL•E, MidJourney and similar software solutions. 

During the presentation intended for journalists, Harris and Raskin demonstrated the way in which, based on a simple textual input, generative artificial intelligence was able to create never before seen photo-realistic content. 

At first, the people nodded in approval as the exclamations of amazement resounded throughout the auditorium – “such a cool thing, very exciting”. 

Then, following the demonstration and the initial strong impressions, the interested attendees nevertheless asked “wait a minute, how could this be – there must certainly be a database from which these images have been drawn?” The audience simply could not believe that the acquired visual was not an already existing photograph, but was indeed created “out of thin air”, based on a simple textual demand. 

Had someone asked me ten years ago whether we would be able to create text-based images, I would have said “No, no way this may happen”.
Slobodan Marković, digital advisor, UNDP and speaker at the “World Ahead” conference

The minds of the journalists, that had expanded due to new information and amazement, slowly shrunk to their earlier state, as people were left somewhat bewildered (and worried?) in the face of something new and certainly revolutionary. 

This is the very phenomenon that we call the rubber band effect.

At one point our cognitive ability to understand something entirely different from what we have already known virtually implodes, as we soon return to our earlier position, a little bit stung with the earlier revelation. 

However, unlike the journalists who had listened to Harris and Raskin, the participants of our “World Ahead” conference found out that they could use their cerebral rubber bands in the best possible way – and with the help of the artificial intelligence (AI) in question. 

We learned that there is a history teacher from Požega who uses ChatGTP to imitate certain historical figures, such as Dušan the Mighty, where pupils can ask questions and thus approach their classes somewhat differently.

The teachers emphasise that some of them allow their students to use ChatGTP during their classes and improve their already written papers, ensued by an analysis of the applied improvements in writing. 

This is a good example showing that, if we manage to identify new ways in which (generative) AI could help us be more precise and efficient, delegating it routine pedestrian tasks — we may leave our fear of being replaced by AI aside, at least for now. 

However, does the humanity going through a poly-crisis, need to approach the challenges of artificial intelligence “through a systemic and continuous building of competences”, or, as we have already heard on several occasions during the conference, through the concept of lifelong learning

Are we, with all these challenges, crises, climate changes and armed conflicts, on the cusp of the sixth mass extinction, or indeed a renaissance is about to come? What door we may choose to open today depends on us entirely. 

The interdisciplinary approach to education is no longer put into question.
Nebojša Vasiljević, director of the Petlja Foundation and speaker at the “World Ahead” conference

As mentioned by Draško Drašković, Head of Exploration in the UNDP Accelerator Lab in his conversation with Katarina Anđelković of the Petlja Foundation, according to some estimates, in the years to come, AI would become our partner in increasing productivity by up to 30%, while the quality of output may be improved by as much as 40%.

Draško thus emphasised that AI may become our path towards the four-day work week. “And we are all looking forward to that” added Katarina, voicing the opinion of many people in the audience. 

If in the last year, just like many people all over the planet, you have also wondered at least once whether robots would take our jobs, or actually replace our friends and romantic partners – we can offer some comfort and encouragement. 

I have this example of my own – I wouldn’t want a chatbot to inform me about my diagnosis, but rather a learned physician. As humans, we need to be facing someone compassionate, as much as proficient, who won’t hesitate to help us, but would also offer support and different options. 

However, AI may be crucial in preventing breast cancer. The Project titled Application of AI in Mammography has been developed to accelerate the process of mammographic diagnostics and thus decrease the time required for expert interpretation of the results, but also reduce the possibility of human error. This system implemented by the Institute for Artificial Intelligence and Institute for Oncology and Radiology of Serbia, with the support by UNDP and the Ministry of Health, could also predict an increased risk from illness, which apparently often remains below the radar of the professionals and experts. 

I believe that the right path involves combining integration of AI into human development and maximising the potential of the AI. On the other hand, we should not neglect the human factor in every phase of the development and supervision of AI. 


An example of such complementary and inclusive use of AI is the software developed in Serbia, which translates speech into sign language by means of artificial intelligence. The SignAvatar has been devised to help persons with hearing impairment in their everyday life, and it is currently used to communicate information concerning the train schedule at the Prokop train station in Belgrade. 

Though the ChatGPT has been a part of our reality for (merely?) one year now, some professors at the Faculty of Dramatic Arts and Faculty of Philosophy, such as Ana Martinoli and Aleksandar Baucal, have been striving to keep in touch with the academic trends where AI tools are used in generating text, image and sound. In addition to using them in their research, students at their exams need to provide a clear and detailed explanation of the way they used and implemented the aforementioned AI tools. 

The high school students who participated at a mini presentation/exhibition 10 selected teams #STEAMchallenge during the conference, have shown that they clearly embrace interdisciplinary learning and complementary use of artificial intelligence. 

I would highlight but a few examples aimed at solving local problems with the application of STE(A)M disciplines, digital technologies and data, such as TechnoKraguji, made by students of the First Kragujevac Grammar School, attempting to solve the problem of pollen fluctuation and the related allergic reactions, as well as the general deterioration of the quality of life. This solution involves mounting pollen concentration tracking sensors, accompanied with developing web and mobile apps. 

And if we consider the calculation made by Professor Baucal at the panel entitled “Does AI Change the Way We Learn?” that the children born in this year would join the labour market in 2046 – we may be facing a bright future that we nevertheless should already fight and advocate for today, together with any given type of artificial intelligence. 

Just as we all use e-mail today as an indispensable method of communication, the generative modes of AI would also become an unquestionable part of our daily and professional routine. Until then, we should educate those who educate, as it has been emphasised by prof. dr. Dubravko Ćulibrk, director of the Research and Development Institute for Artificial Intelligence. 



World Ahead conference

Prof. Ana Martinoli, participant of the panel “Responsible and Critical Use of AI” also referenced creation which, compared to artificial intelligence, is still a unique union of human empathy, intuition, talent or capability, intelligence and all human senses, but also the ability to see a bigger picture and identify local, as well as historical context.

In this sense, the answer obtained through practice concerning whether the “real” intelligence and creativity are lagging behind the “artificial” one, is that we are still in the safe zone, but that our estimates for the following period must not be rigid. 

In the words of David Lynch, the future is so bright that we need to wear dark glasses. The world ahead is truly bedazzling at times, but it also involves hope. Aleksa Gordić developed YugoGPT, a Serbian version of generative AI in the vein of ChatGTP, while in the Science Fund, a project titled Tesla has been approved to deal with the development of linguistic technologies in our language. 


Artificial intelligence would not conquer us as long as we are omnipresent in its development and supervision. Children do not lack digital skills, but we are nevertheless aware that we from the different domains of education must all work together on their knowledge and the related formation of opinions and attitudes.