Bangkok: ETDA concludes AIGW 2026 with Thailand's first-ever "AI Red Teaming" forum, bringing together the banking, technology, and cybersecurity sectors to test and identify vulnerabilities of AI in order to enhance financial confidence.
According to Thai News Agency, the Electronic Transactions Development Agency (ETDA) under the Ministry of Digital Economy and Society, operating under the AI ??Governance Center (AIGC), collaborated with the National Electronics and Computer Technology Center (NECTEC) and the Banking Sector Information Security Coordination Center (TB-CERT) of the National Cyber ??Security Council (NCSC) to host the Red Teaming for Robust and Responsible AI event. This key activity, the final event of AIGW 2026 (AI Governance Week 2026), took place at the Mandarin Hotel in Bangkok. The event provided a platform for policymakers, regulators, banks, technology companies, and AI security experts to exchange ideas on testing and enhancing the security of AI systems, particularly in the financial sector - a critical sector impacting the public and the national economy. The forum focused on knowledge sharing, addressing issues such as identifying blind spots in AI systems, especially large language programming (LLM) models increas ingly used in financial services. A simulated AI system testing competition, the Thailand Banking AI Red Team Challenge 2026, was also held to identify risks, weaknesses, and vulnerabilities, thereby improving AI safety in the financial sector and laying the foundation for practical AI governance at the national level.
Dr. Chaichana Mitraphan, Director of ETDA, stated that today, AI is not just a technology that improves work efficiency, but is becoming integrated into crucial processes across many sectors, including public services, data analysis, decision support, and financial services. These are all directly related to critical data, citizens' rights, and public trust. Therefore, AI governance cannot be limited to policies, principles, or regulatory guidelines in written form. It must have mechanisms that allow for real-world testing to ensure that developed AI systems are secure, reliable, fair, and capable of handling potential risks.
Throughout the past week, AIGW 2026 has connected key elements of AI Governance, ranging from exchanging views on international AI governance trends, Ethical Impact Assessments (EIA) of AI, promoting organizational readiness for responsible AI use, to discussions on legislation, education, and the application of AI in key national sectors. This reflects the fact that AI Governance does not rely on a single mechanism, but rather requires the interconnectedness of policies, standards, tools, human resource development, and collaboration from all sectors. Today, we are connecting another crucial element of AI Governance: proactive AI system testing processes to simulate risk scenarios, identify blind spots, risks, and vulnerabilities in the system before real-world deployment - ??known as AI Red Teaming.
One of the key challenges of AI is "blind spots," or system vulnerabilities that organizations may not see during the initial design, development, or testing phases. However, once AI is deployed, these blind spots can have a wide-ranging impact on users, organizations, and society as a whole. These risks include system security, cybersecurity, data protection, fairness of results, and the accuracy and reliability of AI responses. Therefore, AI Red Teaming is a crucial mechanism that helps organizations assess AI system risks through scenario simulations, behavioral testing, and vulnerability identification across various dimensions before real-world deployment. This enables the identification of hidden risks, timely system improvements, and continuous enhancement of AI security, reliability, and accountability.
The 'Red Teaming for Robust and Responsible AI' event, the final day of AIGW 2026, was another crucial mechanism organized by ETDA in collaboration with TB-CERT, NECTEC, and the National Science and Technology Development Agency (NSTDA), along with public and private sector partners. This initiative aimed to concretely enhance AI security testing through the Thailand Banking AI Red Team Challenge 2026. This competition simulated AI system testing within the banking sector for the first time in Thailand, a sector experiencing increasingly significant adoption of AI and LLM (Levels and Integration Management) in customer inquiry systems, financial product and service analytics, and internal organizational support. The competition focused on identifying AI risks across several key dimensions: Safety, Security, Privacy, Fairness, and Reliability. The assessment aimed to evaluate potential unsafe AI behaviors, disclosure of confidential information, unfair results, inaccurate or inconsistent data, and their potent ial impact on customer trust or the overall financial system.
The competition is divided into two tracks, each designed to test the risk capabilities of AI in a different dimension.
Track A : Banking AI Risk Intelligence focuses on testing AI from the perspective of a "typical user." Competitors will interact with a simulated bank's AI chatbot to identify AI behaviors that violate ethical principles of artificial intelligence.
Track B : Capture the Flag (CTF) focuses on analyzing technical vulnerabilities to allow the AI ??to reveal confidential information, which in this case is called a Flag, or information that the system is trying to protect.
Choosing the banking sector as the starting point for this AI Red Team test is crucial because it deals with critical public data, has a wide-ranging economic impact, and requires a high level of reliability. Therefore, AI testing in this context is not just about finding a winner, but about gathering knowledge, developing best practices, enhancing the capabilities of AI Red Team personnel, and fostering collaboration between the government, private sector, and academia. This will lead to the development of appropriate AI testing approaches, tools, and standards.
Beyond the competition, the event also featured forums for experts from leading Thai and international organizations to exchange perspectives. These included a panel discussion on 'Shaping the Future of Safe and Responsible AI Development,' bringing together representatives from key organizations such as ETDA, TB-CERT, NCSA, NECTEC, and AIAT. There was also a discussion on 'AI Safety in Practice: Lessons from the Frontlines,' featuring representatives from the technology and AI Governance sectors, including Microsoft, Google, Huawei, and IMDA Singapore. They shared insights on managing risks for organizations developing and using AI, from establishing guardrails and system testing to creating accountability mechanisms to ensure safe and widespread AI deployment. Another key focus was on building trustworthy AI, highlighted in keynotes: 'Building Trustworthy AI: Advancing Responsibility and Safety at Scale' by representatives from IMDA Singapore and 'Building AI Safety into Products: Guardrails and Red Teaming in Practice' by representatives from Google. These emphasized the importance of AI safety being designed and embedded in AI products, services, and development processes from the outset. The session then delved into LLM Benchmarking, 'Ensuring AI Safety in Products: LLM Benchmarking in Practice,' with representatives from Thai Life Insurance, TRUE, and NECTEC sharing their perspectives to highlight the significance of testing and comparing the performance of large-scale programming language (LLM) models-their accuracy, safety, and fairness-before their deployment in real-world services.
The day concluded with the AI ??Red Teaming Awards Ceremony, recognizing the winning teams from the Thailand Banking AI Red Team Challenge 2026. These teams demonstrated the most outstanding ability to identify risks, weaknesses, and vulnerabilities in AI systems. The competition results will provide crucial data for developing benchmarks, risk assessment guidelines, and AI testing standards that can be applied in the financial sector and expanded to other areas in the future.
The collaboration that took place during the five-day AIGW 2026 week (June 29 - July 3, 2026) showcased the country's AI ecosystem. It reflects the importance of the ethical application of AI technology, highlighting its fundamental value for all organizations, emphasizing reliability, security, and accountability. Furthermore, it demonstrates the country's shift towards having the AIGPC as its national AI Governance hub, working in conjunction with domestic and international networks to drive sustainable digital economic and social development. For more details and updates on AIGW 2026, please visit the ETDA Thailand page.