When Luis Miguel Trinidad began studying engineering at the University of Seville four years ago, terms such as ChatGPT, Gemini, Claude, prompts, and tokens were unfamiliar to most university students. By the time he graduated this summer, they had become part of everyday academic language, reflecting a transformation that reshaped not only how students completed assignments but also how they learned, collaborated, and interacted with their professors.
As reported by El País, this year’s graduates represent the first cohort to complete an entire university degree during the rise and widespread adoption of generative artificial intelligence. Their experience has become a case study in how rapidly evolving technology has altered higher education, offering unprecedented academic assistance while raising difficult questions about critical thinking, assessment, and the role of universities themselves.
Now preparing to enter the job market, Trinidad reflects on a university experience fundamentally changed by AI. The tools helped him navigate difficult courses more efficiently and organize complex information, making academic progression considerably easier. Yet he believes those advantages came at a cost. Looking back on his degree, he says something fundamental has changed within universities themselves.
That sentiment is echoed by many students and educators who have spent the past several years adapting to technology capable of producing essays, solving advanced mathematical problems, generating computer code, and answering complex academic questions in seconds. While generative AI has become deeply integrated into university life, institutions continue to wrestle with how to preserve meaningful learning in an environment where many traditional assignments can now be completed with minimal effort.
The adoption has been remarkably rapid. According to a survey by Spain’s Knowledge and Development Foundation (CYD), cited by El País, 89 percent of students at Spanish universities now regularly use generative AI chatbots. A separate international survey conducted by education platform Chegg found that four out of five university students across 15 countries use similar tools during their studies. Eurostat data also place young Spaniards among Europe’s most active users of generative AI, with 75 percent of people aged 16 to 24 reporting use of the technology in 2025.
Students say they rely on AI for a wide range of academic tasks. Many use chatbots to clarify difficult concepts or gather information for research, while others employ them to draft essays, solve programming challenges, or organize study materials. Some have gone further by subscribing to premium versions that provide faster responses, more sophisticated reasoning, and fewer usage limits.
Emerging research, however, suggests that improved academic performance may not always reflect deeper understanding. A study conducted by the Centre for Economic Policy Research followed more than 26,000 secondary school students in China over a period of 30 months. Researchers found that after students began using AI chatbots, homework grades increased by approximately 18 percent while the time required to complete assignments declined by nearly one-third. Yet when students sat supervised examinations without AI assistance, their scores fell by roughly 20 percent.
Many students interviewed by El País recognized this contradiction. Trinidad explained that while AI accelerated his progress through difficult subjects, faster completion did not necessarily translate into greater comprehension. Rather than replacing learning entirely, he suggested, many students increasingly delegate significant portions of the learning process itself to conversational assistants.
Scientific research has also begun exploring the technology’s cognitive effects. A study conducted by researchers at the MIT Media Lab monitored the brain activity of university students using electroencephalography while they wrote essays. According to the findings cited by El País, students relying on ChatGPT displayed substantially weaker neural connectivity than participants completing the same task independently.
The influence of AI has extended beyond learning into academic integrity. Esther Valladolid, who also graduated in engineering from the University of Seville this year, said many students increasingly prioritize passing courses over mastering the material. She acknowledged that assignments often appear polished and sophisticated even when their authors cannot fully explain the reasoning behind their own submissions.
At the same time, new technologies designed to disguise AI-generated writing have complicated efforts to detect misuse. Álvaro Muñoz, a recent business administration graduate from the University of Valencia, described the growing use of so-called AI “humanizers” that rewrite chatbot-generated text to resemble natural human writing. The result, he said, is an emerging cycle in which artificial intelligence is increasingly being used to conceal the involvement of artificial intelligence.
Educators report additional challenges inside the classroom. Cristian Castillo, a business administration professor at the Open University of Catalonia, recalled a student who replaced his own correct examination answers with incorrect responses generated by ChatGPT, illustrating what he sees as growing overreliance on the technology even when students possess the necessary knowledge themselves.
For Sandra Calles, who recently completed a degree in education at the Autonomous University of Madrid, AI has also changed students’ expectations about academic effort. She observed that many become frustrated when conversational models fail to provide correct answers immediately, reflecting a shift in attitudes toward the time and persistence traditionally associated with learning.
Faculty members say the disruption extends well beyond coursework. Francesc Pujol, professor at the University of Navarra and founder of the university’s internal AI training platform EurekAI, believes many lecturers feel overwhelmed by the pace of technological change. He argues that universities are simultaneously confronting the lingering effects of the COVID-19 pandemic on students’ education while facing widespread anxiety among young people about artificial intelligence’s potential impact on future employment.
Several graduates also described changes in university social life. Muñoz said classmates increasingly rely on AI instead of gathering to solve difficult academic problems together, reducing opportunities for collaboration outside the classroom. Others believe the relationship between students and professors has weakened as AI increasingly competes with traditional teaching. Diana Barceló, a graduate in labor relations from the University of Alicante, suggested that in some cases students perceive language models as offering clearer explanations than classroom lectures, potentially contributing to declining attendance.
Tiara Lomba, who graduated in psychology from the University of Seville, described what she called a contagious pattern of AI use. Initially turning to ChatGPT during a particularly stressful period of heavy coursework, she found it increasingly difficult to stop relying on the technology, especially as classmates who used it extensively often achieved stronger academic results.
Students have also highlighted growing inequalities linked to access. Premium subscriptions, which can cost up to €100 per month, provide significantly greater capabilities than free versions, raising concerns that financial resources increasingly influence access to more powerful educational tools.
Atman Moussaoui, who recently completed degrees in both German Philology and Classical Philology after eight years at university, believes institutions themselves share responsibility for the situation. Speaking to El País, he argued that universities have failed to provide sufficient guidance on the ethical and responsible use of AI. According to CYD Foundation data, only 34 percent of Spanish university students report receiving formal instruction on AI from their institutions, while nearly half say they would like such training but have not received it.
University leaders increasingly acknowledge that adapting to AI presents structural challenges. María del Carmen Romero, director of the School of Computer Engineering at the University of Seville, noted that public universities often struggle to respond quickly because of institutional bureaucracy and slower decision-making processes.
Many educators now believe the debate has shifted away from detecting plagiarism toward redesigning higher education itself. Pujol argues that by 2025 universities largely accepted that identifying AI-generated work was no longer the central issue. Instead, institutions began focusing on protecting genuine learning in an academic environment where sophisticated conversational models are likely to remain permanent.
Some professors advocate a return to assessment methods that emphasize direct human interaction. Luz Rodríguez, a professor of labor law at the University of Castilla-La Mancha with more than three decades of teaching experience, believes universities should restore greater emphasis on oral examinations, debate, classroom discussion, and supervised activities that require students to demonstrate knowledge in real time rather than through written assignments completed outside the classroom.
Others believe artificial intelligence should be deliberately excluded from certain courses, particularly those taught during students’ first year of university. For Romero, responsible integration rather than complete rejection offers the most realistic path forward. As generative AI continues to evolve at a rapid pace, she argues, universities face the challenge not only of teaching students how to use the technology, but also of ensuring it enhances rather than replaces the learning process at the heart of higher education.
