AI moves beyond coding to strengthen language skills in technical education

Technical skills alone are no longer enough to prepare graduates for engineering and science careers. Employers increasingly report that STEM graduates struggle to write reports, communicate research, and collaborate across disciplines, creating a gap between technical competence and professional performance that universities are under pressure to close.

A new study published in Education Sciences examines this gap. Titled AI-Enhanced Digital STEM Language Learning in Technical Education, the research analyzes whether artificial intelligence–supported digital learning environments can systematically strengthen scientific and professional language skills among STEM students.

Why language has become a critical weakness in STEM education

STEM programs traditionally prioritize mathematics, engineering design, and scientific theory, while communication skills are treated as secondary or outsourced to general language courses. This separation, the authors argue, leaves students ill-prepared for professional environments where writing technical documentation, presenting research, and collaborating across borders are routine requirements.

According to the study, scientific language competence extends beyond general English or Russian proficiency. It includes the ability to interpret academic texts, apply discipline-specific terminology, structure research papers, describe experimental methods, analyze data linguistically, and communicate ethically within scientific norms. These skills are rarely integrated systematically into STEM instruction.

Digitalization and AI, the authors suggest, create an opportunity to rethink this separation. Adaptive platforms can personalize learning, support self-paced progression, and provide continuous feedback in ways that traditional classroom instruction struggles to match. However, most existing digital language tools are generic and insufficiently aligned with the specific discourse practices of engineering and science.

To address this gap, the study proposes a structured AI-enhanced framework tailored explicitly to STEM language learning, combining digital pedagogy with technical communication objectives.

Inside the AI-enhanced framework and classroom experiment

The authors design their instructional model using the ADDIE framework, a widely used systems approach to educational design. The model integrates analysis of STEM communication needs, structured digital content development, guided implementation, and continuous evaluation. Rather than focusing on conversational language, the framework centers on scientific discourse and professional communication.

The digital learning environment developed for the study includes eight instructional modules covering core areas of STEM language use. These modules address academic research skills, technical writing, terminology acquisition, visualization and presentation of data, language for engineering design and calculations, interpretation of scientific results, ethical citation practices, and reflective self-assessment. Artificial intelligence elements are used to support adaptive learning paths, automated testing, and progress monitoring.

To test the framework, the researchers conducted a quasi-experimental study with 122 first-year STEM students from the Mining Faculty at Abylkas Saginov Karaganda Technical University. Students were divided into experimental and control groups. The experimental groups used the AI-enhanced digital platform for scientific language instruction in English and Russian, while control groups followed conventional teaching methods without the integrated digital framework.

Student performance was measured before and after the intervention using standardized assessments aligned with the eight modules. Because the data did not follow a normal distribution, the researchers applied non-parametric statistical methods to evaluate learning outcomes. Reliability testing indicated high internal consistency in the assessment instruments.

The results show that students in the experimental groups achieved significantly higher gains across nearly all assessed areas compared with the control groups. Improvements were particularly notable in academic writing, research communication, terminology usage, and the ability to structure and present technical information clearly. The findings suggest that AI-supported digital environments can accelerate the development of scientific language skills when instruction is closely aligned with disciplinary needs.

What the findings mean for universities and workforce readiness

As automation reshapes technical roles, communication skills are increasingly cited as a differentiator between routine technical work and higher-value professional tasks. Engineers and scientists are expected not only to solve problems but also to explain solutions, document processes, and collaborate across cultural and institutional boundaries.

The authors argue that treating language competence as a standalone subject risks marginalizing it within STEM education. Instead, AI-enhanced platforms allow communication training to be embedded directly into technical learning pathways. This integration supports continuous practice and contextualized learning rather than isolated coursework.

The study also highlights scalability as a key advantage. Digital platforms can support large cohorts of students with limited instructional resources, a growing concern in technical universities facing enrollment pressure. Adaptive systems reduce reliance on one-size-fits-all instruction and allow learners to progress according to individual needs while maintaining standardized learning objectives.

However, the authors caution that technology alone is not sufficient. Effective implementation requires pedagogical alignment, faculty training, and institutional commitment. AI tools must be designed to support learning objectives rather than replace human instruction. The framework presented in the study emphasizes the complementary role of educators in guiding, contextualizing, and assessing student progress.