needs assessments, there is a risk that AI-generated content may subtly guide policy directions or resource allocation without adequate scrutiny or local validation. When GenAI systems generate policy recommendations, funding priorities, or targeting models, the opacity of their training data and algorithms can obscure embedded biases, political assumptions, or methodological flaws. In practice, this could result in skewed prioritization. More importantly, the affected communities may lack the technical or institutional means to identify, question, or contest these AI-generated proposals, particularly when wrapped in technocratic language. Without transparent documentation of how GenAI outputs are produced, and without clear pathways for community feedback or correction, these tools risk disempowering the very populations development aims to serve. This section uses Thailand’s attempts to incorporate and govern emerging technologies in its development initiatives as illustrative cases as to why development agendas need to be grounded in the principle of transparency, accountability, and explainability. As it stands, Thailand's current approach on AI governance, while influenced by the EU’s AI Act, lacks the same depth and enforceability concerning GenAI. Under Article 52 of the Act, the EU requires that any content generated or manipulated using AI, be it deepfake images, audio, or video, must include clear, visible labelling to inform users that the content is artificially produced. Deployers of GenAI systems are also expected to disclose AI involvement and apply technical safeguards to embed traceability into AI-generated content (European Parliament 2023). On the other hand, Thai legislation does not include any requirement for clear labelling of GenAI-generated content. There are no mandates for transparency or traceability, and in this manner, the burden of detection often falls to users, journalists, or civil society, who lack the tools and resources necessary to respond at scale. Equally important is the role of social media platforms, which serve as primary channels for the distribution of GenAI content. The UK’s Online Safety Act of 2023 exemplifies a state’s attempt to hold platforms at a higher standard of responsibility in moderating manipulated media. Algorithms that amplify engagement over accuracy can accelerate the spread of misinformation; thus, platforms should be required to disclose when and how AI is used in content generation or recommendation engines, and to deploy detection tools that identify and flag inauthentic content in real-time (Richards 2025). Greater transparency through regular audits and clearer content labelling would help restore public trust and support efforts to build a safer digital environment. Overall, the absence of explicit provisions on 6