TMA Joins OSBC 2025 in Seoul: Advancing Responsible AI in the Global Open-Source Community

On the occasion of Vietnam’s National Day (September 2, 2025), TMA Solutions joined hands with Insignary and OSBC to successfully co-organize the 14th OSBC Open Source Conference in Seoul. With the theme “Connecting AI in Open Source & Software Supply Chain Security (SBOM)”, the event brought together hundreds of experts, developers, and technology leaders from around the world to discuss emerging trends in artificial intelligence, open-source software, and cybersecurity.

As a prestigious annual event of the global open-source community, OSBC not only delivered a series of in-depth sessions but also created valuable opportunities for knowledge-sharing and collaboration among technology enterprises and international developers. This year’s focus addressed critical industry challenges, including Artificial Intelligence (AI), Software Bill of Materials (SBOM), and the EU Cyber Resilience Act (CRA).

Representing TMA Solutions, Mr. Dung Dao – Director of the AI Center, was invited to deliver a presentation titled “Responsible AI in Open-Source AI Models: Challenges and Solutions.”

In his presentation, Mr. Dung highlighted both the opportunities and risks of open-source AI models. While adoption is rapidly increasing due to lower costs, faster time to value, and strong developer interest, organizations face critical challenges related to cybersecurity, compliance, and intellectual property. He introduced the concept of Responsible AI, which is grounded in fairness, transparency, accountability, privacy, and reliability. These are essential principles to guide the safe and sustainable use of AI.

Mr. Dung also shared practical solutions to strengthen AI accountability, including the use of model cards to build trust, AI Bills of Materials (AI BOM) to document design and dependencies, and MLOps pipelines to ensure reproducibility and audit readiness. He further explained approaches to reduce bias in both traditional machine learning and generative AI, and tools to mitigate hallucination in large language models.

Drawing on TMA’s own experience, the session included case studies of how Responsible AI has been applied in customer projects. These ranged from safety services for wearable devices and smart agriculture solutions (T-Pest) to telehealth knowledge-based chatbots. Each case demonstrated how Responsible AI practices, from bias assessment to evaluation frameworks, help TMA deliver solutions that are both innovative and trustworthy.

Highlighted case studies applied AI of TMA Solutions:

• T- Pest Smart Agriculture solutions with AI, IoT, and GIS
• Enhancing Healthcare with a Virtual Assistant

By joining the OSBC Open-Source Conference 2025, TMA Solutions reaffirmed its commitment to advancing Responsible AI. Through continuous investment in research and global collaboration, TMA aims to help enterprises harness the power of AI while ensuring safety, accountability, and long-term value creation.