ITU student: AI is everywhere at university, but where is the critical conversation?
Herbert Hertz van den Berk, a third-year student at the IT University of Copenhagen (ITU), argues in an opinion piece for Ingeniøren that AI is everywhere at university but the critical conversation about it is missing. He started at ITU in 2024, expecting a place that would debate new technologies, large language models (LLMs) included, for better and worse. Instead he finds that debate hard to join.
He describes LLMs as present throughout his studies. Students sense that AI skills will be decisive for entering the job market, and the industry seems sold on the technology. AI already shapes how students study, write and think. Around exams, he and his peers discuss whether they feel real ownership of what they write or design when AI has been involved in nearly the whole process.
He sees a split between two worlds. In courses on the subject, AI is taken apart: bias in datasets, model limitations, hallucinations, who is behind the technology, and its social and climate effects. Models are described as impressive but fallible statistical systems that depend on huge amounts of data and computing power. Outside the university, tech industry talk centres on what models might become, including AGI, AI at or above human level that can no longer be controlled. The disagreement rests on whether current models are just advanced statistics or the beginning of something far more intelligent.
He notes that companies like OpenAI and Anthropic have financial interests in presenting their technology as world-changing, so the AGI story could be dismissed as marketing. But similar worries come from researchers and some of AI's "founding fathers," so it is unclear whether superintelligence is hype, a genuine scientific concern or something in between.
His point is that universities should not declare AI good or bad but should be a place to discuss these questions. Without that, each student must navigate between industry promises, media doom stories and academic scepticism. Students may also accept the premise that more AI is necessary before asking whether it is desirable, and may end up learning to use AI rather than to think for themselves.