Resilience Training in Higher Education: AI-Assisted Collaborative Learning

We propose a collaborative hybrid (human plus artificial intelligence) learning, which enhances learning impact with additional elements. These elements are: collective intelligence, created by integrating large language models (LLMs) into argumentation to strengthen students in various roles; adversarial learning, where competition between students is built on the principles of generative adversarial networks around complex decision objectives (dilemmas) towards developing resilience skills. We conduct an experiment, structured as a series of intellectual sparring between teams of players. The scenario of sparring evolved from classical disputes to technologically augmented competitions. The core hypotheses, "positive impact of adversarial training" and "game-changer role of LLMs in argumentation", have been confirmed. An important conclusion is that the added value of LLM tools strongly depends on how professionally they are used. Achieved results are presented as a contribution to collaborative hybrid learning using artificial intelligence as a personal digital assistant.