Page 54 - Social Enterprise A New Business Paradigm for Thailand

For the purposes of budget analysis and other quantitative assessments, the study adhered to
the legal definition of “social enterprise” as recognized under Thai law. This approach was feasible

because the population of legally registered social enterprises is clearly defined. It allowed the
research to address Research Questions 4 and 5, as well as the fifth objective of this study.

1.10 Research Methodology

This study adopted a mixed-methods approach, comprising the following components:

1. Document Review Using Search Engines and AI Tools
The research team conducted an extensive literature review utilizing a combination of search
engines and artificial intelligence (AI) tools. Multiple AI platforms, such as ChatGPT, Bing, and
Perplexity, were used in parallel. Their outputs were then verified using traditional search engines
like Google to ensure source credibility. All gathered information was subsequently read, analyzed,
synthesized, and integrated by the researchers. Notably, AI-generated responses were never
copied directly into the report, i.e., AI is merely a research assistant not a researcher.

For data regarding social organizations, publicly available content was sourced from their official
websites. Since these sites are regularly updated and easily accessible, citation timestamps were
not included in the references. During the review, the research team experimentally tested
multiple AI tools to evaluate their accuracy and reliability. It became evident that AI could present
fabricated or unverifiable content, which led to the development of a formal AI usage protocol
comprising four key principles: (1) clarity of purpose in prompts, where AI prompts must clearly
indicate that the information is for academic purposes to ensure that responses are verifiable and
properly sourced; (2) verification of sources, requiring that any source cited by AI be
independently retrieved by a research assistant using a search engine to confirm its existence; (3)

cross-validation across platforms, in which information obtained through AI, such as summaries
or literature reviews, must be cross-checked using multiple AI tools to identify inconsistencies or
divergent interpretations, with the process modeled after the triangulation method commonly
used in qualitative interviews; and (4) human oversight and synthesis, meaning that all AI-
generated outputs must be validated by directly reviewing the original sources, with researchers
conducting their own independent analysis and integrating findings with primary data and other
materials. In summary, the research team concluded that AI should serve only as a research

assistant, not as a substitute for human researchers. All critical interpretation, synthesis, and
judgment must remain the responsibility of the researchers.

2. In-Depth Interviews
To gather qualitative insights into social businesses, the team applied a broad definition of “social
business”. This inclusive scope allowed for a more comprehensive understanding of the

21