Are ChatGPT-Like Applications Harmful for the Learning Process?

ChatGPT took the world by storm and generated extensive discussions about its uncanny capacity to generate realistic outputs that, superficially, could look like they are written by humans. Especially education appeared it could be hit hardest, due, mainly to slow but sustained evolution of assessment policies in universities and the slow response to perceived threats. This paper analyses the interference between AI creations and their possible effects on the learning processes. It then presents the proposed computer aided learning framework developed and implemented by one of the authors and its capacity to negate the harms that could be generated by ChatGPT applications. This is shown in an experiment. Cybertrainer, the proposed professional learning service has the capability to remove the threat to academic integrity by several special features: purely student-centred learning, full autonomy and freedom to intervene when desired, anonymised interventions, traceability by entirely written contributions, and, critically, robust quality assurance system guaranteeing integrity. Cybertrainer can truly counteract the dangers of ChatGPT by encouraging humans (students) to do what humans do better – critical analysis, synthesis, and creativity.