Artificial intelligence as a catalyst for advancing soft skills in higher technical education

Abstract

Russia`s full-scale invasion has transformed Ukrainian lecture halls into hybrid shelters and digital studios, forcing universities to reassess how they cultivate the “human infrastructure” of future engineers. This study designs and field-tests a low-bandwidth, artificial-intelligence–supported instructional framework aimed at accelerating soft-skill growth – communication, collaboration, critical thinking, adaptive leadership, and digital civic responsibility – under wartime conditions. Over six months (October 2024 – March 2025) the framework was embedded in three power-engineering courses and reached 127 third- and fourth-year students in Vinnytsia National Technical University. Interventions combined (i) chatbot-mediated crisis-negotiation labs, (ii) an edge-AI empathy dashboard that visualised speaking-time equality and affect congruence during Scrum meetings, (iii) deepfake-detection simulations using prompt-engineered language models, and (iv) an offline GPT-Lite writing assistant optimised for blackout periods. Mixed-methods evaluation coupled competency rubrics, multimodal learning analytics, and reflective journals. Negotiation scores rose 3.1 → 4.2 / 5 (d = 0.86); the Gini coefficient of speaking-time distribution fell 0.37 → 0.21 while peer-rated psychological safety climbed 4.9 → 6.1; Ennis–Weir critical-thinking scores improved by seven points; and proposal-revision cycles shortened by 34 %. All gains were statistically significant (p < 0.001) despite frequent power cuts and intermittent internet. Ethical safeguards – data minimisation, encryption, and explicit consent – ensured full compliance with national privacy regulations; raw logs remain confidential and anonymized. The findings demonstrate that context-aware AI can act as a “digital teammate,” externalising metacognitive cues and providing rapid, adaptive feedback unattainable by individual instructors under siege. Because the tools rely on open-source libraries and run on modest hardware, the approach is readily portable to other Ukrainian institutions and conflict-affected regions. Future work should trace graduates into reconstruction projects, refine ultra-lightweight analytics for off-grid operation, and co-create domain-specific modules with industry partners. In sum, AI, when ethically integrated, offers a scalable catalyst for rebuilding not only laboratories but the socio-emotional fabric essential to Ukraine`s recovery and global workforce integration.