The frustration: in the lecture hall, year after year, we see that students use artificial intelligence tools ever more frequently, but with an ever poorer understanding of the mechanisms, limits and ethical implications of this use. Generative AI has democratised access to knowledge, but it has equally democratised access to plausible disinformation.
The result is a paradox that we may call the illusion of competence: the student who has used an LLM platform to generate an academic-sounding introduction, but who cannot
The hope: that there is a middle way — that AI can be a remarkable tool of intellectual productivity without replacing critical thinking, that transparency can become a cultural standard in the Romanian academic environment, and that today’s students can become tomorrow’s researchers of integrity precisely because they have understood, early on, what a language model can and cannot do.
Who is this project for? Final-year students preparing their bachelor’s and master’s theses, doctoral students at any stage of their research, and all those who are looking for a frame of reference for the responsible use of AI and who wish to integrate artificial intelligence tools into their academic workflow without compromising scientific integrity.
Taking part in the project does not require advanced technical knowledge of artificial intelligence. No mathematical formulas, neural network architectures or source code are taught.
Instead, you will find a functional understanding of how generative language models operate — enough to use them strategically and to recognise the errors they inevitably produce.
