The artificial intelligence is increasingly entering production and training processes, but its integration is not proceeding at the same speed as its diffusion. This is one of the key messages that emerges from "EDUNext – New Scenarios for Education and Skills in the AI Era”, the new research from the Look4ward Observatory presented in Rome by Understanding St. Paul and by the Strategic Change Research Center “Franco Fontana” of the Luiss Guido Carli University.
A work that tries to answer a now central question: how learning changes when the AI does it become part of cognitive, educational and decision-making processes?
AI is growing in businesses, but the training gap remains wide.
The study data shows an acceleration obvious in the adoption of artificial intelligence technologies: 31% of the companies surveyed have already adopted or are experimenting with AI solutions, up from 19% in 2025. This figure confirms how artificial intelligence in companies is becoming an increasingly structural element of organizational and production processes.
However, this technological evolution is not accompanied by a parallel development of internal skills on AI and digital training. The picture is in fact heterogeneous and signals a growing gap Between innovation and human capital: 85% of companies with AI have launched or planned dedicated training, but only 19% have structured and ongoing programs, while 48% limit themselves to pilot or occasional initiatives. Even more critical is the data according to which 46% of employees have not received training specifically on artificial intelligence, and 44% of companies do not plan to invest in training in the next 12-24 months. This misalignment highlights a structural issue: technology is outpacing the development of human capital.
AI and Learning: Value Depends on Context
One of the most relevant aspects of the EDUNext research concerns the impact of artificial intelligence on cognitive processes and learning models. The empirical analysis conducted on approximately 800 people, mostly students, shows a surprising result: the effectiveness of AI is neither automatic nor uniform, but depends significantly on the complexity of the task.
In low complexity tasks, in fact, the absence of technological tools favors greater involvement, attention and active learning, suggesting that the use of AI is not always a factor for improvementOn the contrary, for highly complex tasks, artificial intelligence proves to be an effective support, capable of reducing cognitive load and improving the quality of decisions.
A key passage of the study emphasizes that “the value of artificial intelligence is not universal, but contingent and depends on how it is integrated into educational and professional processes.”
The EDUNext Genius Model: Technology and Cognitive Autonomy
Starting from these results, the research proposes a new framework: the Ingenious EDUNext model (Generative Ecosystems for New Intelligent Augmented Learning Education). The aim is to design educational systems capable of integrating in a balanced way human and artificial intelligence, avoiding both the substitution of cognitive abilities and an uncritical use of technology.
The model is based on some key principles of augmented education: selective use of artificial intelligence based on task complexity, the centrality of individual cognitive autonomy, the integration of technological, ethical, and critical skills, and the decisive role of pedagogical design in human-machine interaction. This approach shifts the focus from learning automation to "augmented learning," that is, learning boosted but not delegated to technology.
Voices of Research: Between Transformation and Asymmetries
For Elisa Zscopio Marsala, Head of Education Ecosystem and Global Value Programs at Intesa Sanpaolo, "artificial intelligence doesn't replace learning, but transforms it, and its value depends on the quality of the educational models with which it is integrated." This transformation requires new transversal skills and greater collaboration between businesses, universities, and institutions.
Enzo Peruffo, director of the Luiss Research Center in Strategic Change "Franco Fontana," highlights an "asymmetric" dynamic: "The EDUNext research reveals a profoundly asymmetric transformation: artificial intelligence has already entered everyday study and work activities, but its integration into educational, organizational, and decision-making processes is proceeding more slowly." He adds: "The Geniale model arises precisely from this evidence: augmented learning does not coincide with the automation of learning, but with a more conscious design of the interaction between technology and human capabilities."
