Наредни састанак Семинара биће одржан онлајн у среду, 15. октобра 2025. са почетком у 19 часова.
Предавач: Miguel Ángel Conde González, Department Ing. Mecánica, Informática y Aeroespaci, Universidad de León, Spain
Наслов предавања: WHO REALLY KNOWS MORE? AI, LEARNING, AND THE NEW ALGORITHMIC ORDER
Апстракт: Learning Analytics (LA) has long promised to provide empirical evidence to inform educational decision-making, always with the ambitious goal of improving student learning. Yet the current landscape is evolving at an unprecedented pace. Increasingly, learning data is generated and stored outside institutional boundaries–by external platforms, intelligent assistants, AI-powered tutoring systems, and interactive content generators. This reality challenges the traditional scope of LA and blurs the very boundaries of educational analysis.
Such a shift calls for urgent responses from multiple perspectives. We need new standards to capture and reuse data produced beyond institutional ecosystems, while preserving its analytical and ethical value. Learning platforms must move toward genuine personalization, delivering exactly what each learner needs, precisely when they need it. Intelligent tutors capable of intervening in real time during learning tasks are already emerging, further transforming the landscape. Perhaps most disruptively, the rise of Explainable AI (XAI) promises not only to predict and recommend, but also to justify its decisions–albeit in ways that may be more persuasive than transparent. This raises a profound epistemological challenge: are humans still interpreting data and making pedagogical choices, or are we quietly handing over this responsibility to opaque yet seemingly efficient systems?
This talk offers a critical reflection on this evolving context, marked by a growing dependence on algorithmic models that often escape pedagogical oversight. We will explore both the risks–such as the erosion of institutional data sovereignty and the dangers of misplaced trust in algorithmic explanations–and the opportunities: radical personalization, broader access to analytical tools, and a shift toward more distributed yet contextualized forms of LA.
Ultimately, one pressing question remains: if AI already knows (or seems to know) more than we do, what is the new role of Learning Analytics–and what is ours, as educators and researchers?
Напомена: Семинар је могуће пратити на даљину. Све информације су доступне на страници:
https://miteam.mi.sanu.ac.rs/asset/DPP9i2jhvYzp8dmRe
