Page 642 - AI for Good Innovate for Impact
P. 642

AI for Good Innovate for Impact



                      lab experiences in a fully interactive 3D virtual environment, while integrating AI-powered
                      instructional support to guide students through experiments, assess performance, and provide
                      real-time, personalized feedback. The platform specifically addresses challenges such as the
                      lack of physical lab infrastructure, shortage of expert instructors, language barriers in technical
                      learning, and disparities in access to STEM resources. The AI instructor functions as a virtual
                      mentor, capable of understanding and responding to voice commands in both Amharic and
                      English, guiding learners through simulations, evaluating circuit designs, and offering adaptive
                      hints based on student input via text, speech, or visual interactions.

                      Built using Unity, the 3D simulation engine models electronic components and behaviors,
                      making the lab accessible from standard computing devices without the need for expensive VR
                      headsets. The AI system includes a speech recognition module that converts spoken Amharic
                      or English into text, and a natural language processing (NLP) and dialogue management
                      system to interpret student intent and determine contextually appropriate AI responses. An AI
                      evaluation module analyzes both circuit graphs and visual layouts to assess design accuracy,
                      while Clarifai is used to manage multimodal data—such as images of circuits, voice inputs,
                      and session metadata—for tagging and continuous learning system improvement. All student
                      interactions are logged with metadata including error types and time-on-task, supporting
                      adaptive feedback loops and reporting.

                      The expected outcomes of the project include significantly improved access to quality STEM
                      education for students in rural or underfunded schools, increased student engagement and
                      retention through interactive and hands-on learning experiences, and enhanced personalization
                      of instruction based on real-time AI analysis. Moreover, the inclusion of Amharic speech
                      recognition promotes linguistic inclusion, and the system enables data-driven insights into
                      student learning outcomes, aiding teachers and institutions in making informed educational
                      decisions.

                      Use Case Status: The use case is part of a large product development.


                      Partners

                      O-SKILLS PLC, through its O-SKILLS Technical Education and Training Centre, serves as a
                      key partner by leading the development of the 3D Virtual Electronics Lab environment. Their
                      role involves designing and implementing the immersive 3D virtual lab space that accurately
                      simulates electronics experiments. In parallel, the AI-driven instructional assistant—responsible
                      for real-time guidance, speech recognition, and performance evaluation—will be developed
                      by AASTU students. O-SKILLS will closely collaborate with the AI development team to ensure
                      smooth integration of the AI instructor into the 3D lab, facilitating effective interaction between
                      the virtual environment and the intelligent tutor. [1]

                      2�2     Benefits of use case

                      The project advances equitable access to technical education by replacing physical lab
                      infrastructure with an AI-driven 3D virtual electronics lab, enabling underserved students to
                      acquire job-relevant skills in electronics and digital literacy. Its adaptive learning paths and
                      real-time feedback bridge educational disparities, ensuring inclusive skill development aligned
                      with STEM employment demands. Multilingual support (e.g., Amharic) and gender-neutral
                      design further dismantle barriers to participation, targeting systemic gaps in technical and
                      vocational training.



                  606
   637   638   639   640   641   642   643   644   645   646   647