This is not a uniquely Cuban problem, but a universal one. Today, however, I want to briefly comment on the characteristics that are very particular to the context where I work as a faculty administrator. As I mentioned, I was recently appointed Vice Dean for Academic Affairs at one of the faculties of my university, and I feel compelled to focus on this complex issue, which unfolds like a hydra with a thousand heads.
The Cuban specificity is explained by the practical need to rely on virtual learning environments as a fundamental resource for the development of the teaching-learning process. In other words, what in normal times functions as support—admittedly, very important support—has now taken on a starring role in the midst of a major crisis marked by a chronic fuel shortage, which in the case of my university has forced us to rethink the ways we organize teaching. Since March, students—who come from all over Cuba—have been sent back to their homes, so they are receiving their classes remotely.
But to speak of "remote classes" without qualification is misleading in the Cuban case, because for many—including professors—access to the internet becomes impossible due to prolonged power outages and the resulting connectivity problems. Under these grim conditions, we are committed to continuing to train computer science professionals across various areas, ranging from cybersecurity to bioinformatics.
My experience tells me that overcoming the challenge posed by generative artificial intelligence is highly complex. That is: how do we ensure that students make ethical, transparent, and therefore lawful use of artificial intelligence when carrying out the various tasks assigned by their professors, thus avoiding abuse of its infinite benefits?
After a considerable time of extensive experimentation with different LLMs, I keep insisting to my colleagues that in a virtual environment it is very difficult to prevent the teaching-learning process from being compromised in this regard. This is especially true because options such as live virtual sessions, synchronous meetings, or timed quizzes are impossible to implement under Cuba's current infrastructural conditions. And so-called AI detectors are increasingly behaving like mere bounty hunters, default-labeling everything as "generated with artificial intelligence" just to gain visibility or promote their paid plans.
From the social sciences to the exact sciences, I see no other path than, ultimately, that oral or practical moment when the student stands before the professor, defending the organic level of knowledge they have achieved—including their level of expertise as prompt engineers.
I repeat: in other contexts, this same issue might be resolved virtually, but in Cuba's current scenario it is highly complex, because many students will never coincide online with their professor, whether due to problems with the electricity service, connectivity issues, or just urgent socioeconomic needs.
I will be sharing here the ideas that I design and implement with my team as we try to manage this entire situation during the course that starts next week. I would be greatly encouraged by any suggestions or ideas that might work for us here.
Source for the cover image, generated with Canva.
