An AI diet: cognitive delegation arrives in the classroom
More than nine in ten students are already using artificial intelligence for schoolwork. We talk to Lorenzo Alberca García, head of Trade Union Action at STE-CLM and coordinator of the Artificial Intelligence Area at STEs-Intersindical, about the macro-study that puts a name to the problem: «cognitive delegation».
More than nine in ten students in this country say they already use artificial intelligence for schoolwork: 92.74% at national level and 90.54% in Castilla-La Mancha. That is what a macro-study by the STEs-Intersindical trade union, carried out in May, tells us, with nearly 6,000 teacher surveys and more than 1,000 from students, across the seventeen regions and the two autonomous cities, from Primary to vocational training. Fran and Carlos chat with its mastermind, Lorenzo Alberca García, head of Trade Union Action at STE-CLM and coordinator of the Artificial Intelligence Area at STEs-i, who admits the figures were what they expected: the study has simply put numbers to what was already suspected. And he warns: AI «is unstoppable, but that does not mean it should be adopted without any kind of review or demands».
Lorenzo explains the gap between classrooms and staff rooms: 86% of teachers know about AI, but only one in four uses it frequently, and the main barrier is work overload. For him, this may be the first tool in history that students adopt before teachers do, something that did not happen «not even with the clay tablets of Sumer, or with books, or with computers». An hour a day of AI for school tasks is joined by another hour of ICT in the classroom, another hour of AI for private use, and three hours on social media.
The heart of the conversation is cognitive delegation. Lorenzo cites the MIT Media Lab study which, using electroencephalograms, measured a «cognitive debt» in people drafting with ChatGPT, and he draws on a powerful image: if the industrial revolution devalued physical strength, AI means a devaluation of knowledge, and bringing it into school without further thought would be introducing a Trojan horse into an institution whose foundation is transmitting that knowledge.
On the 140 million announced by the Government for public educational AI, the union demands data protection for students, in-house models independent of the big tech firms, negotiation with the unions, and that it should never be used to assess teachers; Carlos adds the awkward question of where the machines are to run it. The teacher-student relationship, Lorenzo argues, is essentially emotional and irreplaceable. And as the star measure, disconnect the classrooms: students’ work should be done in class, which means lower class sizes and more staff. With Olé Olé, Def Con Dos and El Aviador Dro as the soundtrack, we are left with one idea: the question is not what the machine can do, but what we do not want to stop doing ourselves.