AI in Advertising, Education, and Academic Evaluation

The most significant change during this period was the retirement of the GPT-4o model in ChatGPT (and some other older models). OpenAI justified the change by citing a focus on newer versions, but the response from some users was surprisingly emotional—partly because some had been using the chatbot as a companion. Some have even dubbed the model’s discontinuation the “Valentine’s Day Massacre.” For educators, this serves as a reminder that replacing or discontinuing a model overnight can bring about significant changes, so it makes sense to design assignments that do not rely on a single version of the tool.

If the first story raised the question of what happens when a feature disappears, the second raises the question of what happens when the UI experience is monetized. OpenAI is beginning to introduce ads into ChatGPT, which adds an extra layer (sponsored content) to the user experience. In an educational context, this directly impacts information literacy: distinguishing between the model’s response, the actual source, and an advertisement, as well as the question of how such environments influence users’ decision-making.

For those of you who also use AI in projects (e.g., as part of educational apps or research prototypes), there is another, less obvious layer of change to consider: the timeline for model deprecations and replacements in the API (application programming interface). OpenAI regularly updates a list indicating which versions are being deprecated and when, which serves as a good reminder that even in educational solutions, it’s important to consider maintenance and the replaceability of components.

In higher education, there is a growing shift away from blanket bans toward specific rules governing the use of AI for individual assignments. Inside Higher Ed reports on an analysis of a large set of course syllabi, which shows that instructors are increasingly specifying when AI is permitted (and for what purposes) and when it is not, while placing greater emphasis on attribution and transparency in its use.

At the same time, in some settings, AI is becoming a standard part of computer science education. The Wall Street Journal reports on a collaboration between Anthropic and the nonprofit organization CodePath, where students use Claude’s tools to learn programming and work on projects in their courses.

In addition to policies and guidelines, there are also concrete examples where students are actually co-creators of AI solutions. Boston University reports on a project in which students collaborated on the development of a chatbot that assists doctors in caring for patients with impaired consciousness. This is a good example of AI use where the focus is on understanding the domain, responsibilities, limitations, and testing—rather than on the automatic generation of responses.

Another institutional example comes from Ohio State University, where they report concrete progress in integrating AI literacy into undergraduate programs. It is particularly important to note that this involves not just general guidelines, but faculty-specific plans, support for instructors, and the development of a subject-specific assistant. This is a useful example of how a university is implementing AI not merely as a tool, but as a systemic competency for students and staff.

Another highly relevant development is the research-and-teaching approach from George Mason University, which is launching a study on the ethical use of AI in higher education. The project involves multiple institutions, tests introductory activities for the responsible use of AI among students, and also announces open educational resources (OER) that will be more widely available upon completion. This is particularly useful for university instructors because these are transferable activities rather than merely declarative guidelines.

At the course design level, Brandeis University stands out; in an article on the course “Automation and Software Development”, they present a somewhat different pedagogical response to the AI wave—with less emphasis on current tools and more on understanding automation, programming principles, and the critical evaluation of technology. This approach serves as a good reminder that incorporating AI into the curriculum does not necessarily mean more tools, but can also entail a more fundamental, conceptual overhaul of course content.

When we consider the shifts in tools, rules, and practices, a significant change becomes apparent precisely in terms of their impact on the pace and scope of work. The Harvard Business Review points out that, in practice, AI often does not reduce the workload but can actually intensify it, increase the pace, raise expectations, and expand the scope of tasks. This is a relevant framework for higher education as well, where AI facilitates the preparation of materials, but expectations for more variations, greater personalization, and more feedback quickly emerge.

While global tools are changing rapidly, a more structural shift is also taking place in education systems, as AI-generated content is being incorporated into official curricula. In the last three days, news has emerged from South Korea, where the Ministry of Education is introducing a government-supported model for introductory AI courses at universities, under which 20 universities will receive funding to develop introductory AI courses (including for non-engineering programs) and to disseminate these models across institutions. This is an important signal, as it marks a shift from individual pilot projects toward a more systematic, nationally oriented introduction of AI literacy in higher education.

Croatian media report that Croatia is introducing an elective course on AI in secondary schools starting this fall. Thus, significant shifts in the integration of AI into education are also taking place close to home. Let’s take a look at what else is happening in this regard here in Slovenia.

During this period, Slovenia also received news of greater strategic interest. The Ministry of Digital Transformation reports on a new UI development cycle, ranging from the continuation of EDIH work to two key pillars (the Competence Center for Artificial Intelligence and the Slovenian Artificial Intelligence Factory). This is particularly important for the academic community because it indicates where development projects, partnerships, and funding will be concentrated in the coming years.

In Slovenia, attention has recently been drawn to a discussion on the “Digitrajno izobraževanje” podcast (MMC RTV SLO), which explores the connection between AI, classroom dynamics, and the (in)security of young people’s well-being. The discussion touches on the use of artificial intelligence among young people, the importance of critically examining it, and its social implications.

With this latest news, we turn to our university environment. The Faculty of Mechanical Engineering at the University of Ljubljana has developed a virtual assistant, accessible on the faculty’s website, designed to answer students’ basic questions related to their studies. Its operation is monitored and guided by a group of faculty staff who regularly update the learning resource from which the virtual assistant draws its knowledge. Behind the scenes, they have prepared a knowledge base with a set of questions and answers, which they continuously update based on user interactions. This is an interesting example of how AI can help provide quick and reliable answers to students’ most frequently asked questions, thereby reducing the workload on the faculty’s administrative and support staff. For more information about this case study, please contact the Associate Dean for Undergraduate Education, Assoc. Prof. Dr. Matevž Zupančič, Matevz.Zupancic@fs.uni-lj.si.

News from the past two weeks shows that AI in education is increasingly being viewed less as a standalone tool and more as an evolving ecosystem in which models, business models, institutional rules, and user expectations are all shifting simultaneously. For the university environment, this means, above all, the need for more thoughtful task design, clear rules of use, and the consistent development of critical judgment, since the stability of these tools cannot be taken for granted, and their role in learning is becoming increasingly embedded in everyday pedagogical and organizational practices.

If you have explored an interesting use of AI in education at your college, in a course, or as part of your professional work, we cordially invite you to share your example with us via the form for collecting examples of best practices. We look forward to reviewing the submissions and including selected contributions in future issues of the AI newsletter so that, together, we can build an overview of useful and thoughtful approaches in the university setting.

Authors: Eva Kern Nanut, Eva Škraba, Maja Kosmač, Sanja Jedrinović Čufer, Mateja Bevčič, University of Ljubljana Center for the use of ICT in pedagogical process