AI challenges Spanish bachelor's thesis: 80% of universities seek curriculum reform

AI challenges Spanish bachelor's thesis: 80% of universities seek curriculum reform

Is a final year project (TFG) worth anything if an artificial intelligence can write it in minutes? This question haunts Spanish university faculties since the rise of AI exposed the flaws in the Bologna Process. A recent study by the CYD Foundation indicates that 80% of universities believe their curricula need modification and 63.9% prioritize adapting assessment methods as the most urgent task. Julio González, Professor of Administrative Law at Complutense University, states bluntly: "Today’s TFGs make no sense. When supervising final projects, the central issue now is spotting where the AI was used."

The TFG written in minutes, defended in hours

The Bologna Process arrived in Spain around 2010, promising a shift from content transmission to competency acquisition. Degrees were shortened to four years, introducing a final project designed to consume between 60 and 120 student work hours. Generative AI completes this same task in an afternoon. The paradox is that a system meant to measure competencies now primarily measures the ability to disguise shortcuts.

González does not ask to ban the tool. He asks to use it as "another knowledge instrument," comparable to a book. His diagnosis is uncomfortable: "Students are not learning what they should because they look for ways to pass rather than to learn." He adds a warning that sounds like a labor verdict: "In the job market, either you provide professional answers or you will disappear."

What says the university that still defends the TFG

Rubén González, Vice-Rector for Organization and Academic Planning at UNIR, disagrees with the diagnosis but agrees on the urgency. For him, Bologna is not obsolete; what fails is how AI is used. UNIR has introduced continuous assessment allowing them to "interrogate the student at any moment" to confirm real knowledge via oral exams. They have also reformulated assignments toward a theoretical-practical approach with reasoning and solutions.

Regarding the TFG, the vice-rector argues that "with good supervision it adds value" and that defense before a tribunal still reveals who worked and who did not. "Facing such a moment, for a 20 or 21-year-old student, is a challenge that forms part of their education." The difference between these positions is significant: one questions the instrument, the other only its use.

Congress acts while debate stalls

The Congress Committee on Science, Innovation and Universities approved an initiative by Sumar to create a space for reflection and shared governance on the responsible integration of AI in the Spanish University System. The measure comes late for those already dealing with the problem in classrooms, but marks a territory: institutional response moves between pedagogy and surveillance.

Researcher Laia Lluch proposes using AI as "an ally that makes us think," a formula that sounds good in headlines but requires something prior: someone teaching students to think with it. In class, notes the professor, this is not happening. "Misuse of AI is the same as copying; this is unregulated and is a problem."

The job market waits for no one

The discussion extends beyond the classroom. In a competitive market, according to González, the difference will be made by "those who know how to use AI like a book, to study, analyze and give answers, versus those who just copy-paste, who go nowhere." UNIR goes further, defining the tool as "a companion that enhances what you can learn," with human supervision as an indispensable condition. Their Psychology students already train with practical cases created by teachers using AI to arrive better prepared for clinical internships.

The underlying problem is incentives. If the system rewards passing rather than learning, AI merely accelerates existing logic. If it rewards demonstrating judgment, AI becomes a lever. Technology decides nothing on its own. Assessment design does.

---

With these elements, it would be reasonable for universities to have reformed assessment two academic years ago. They continue debating whether the TFG makes sense. Meanwhile, AI has already submitted its own.