meta-synthesis

Meta-synthesis is a rigorous qualitative research method for systematically integrating, comparing, and re-interpreting findings from multiple independent qualitative studies to generate new, comprehensive insights. Unlike meta-analysis, which handles quantitative data, it focuses on unstructured text from interviews, case studies, and observations. In risk management, it's a key tool for risk identification and assessment, particularly for socio-technical risks. For instance, the NIST AI Risk Management Framework (AI RMF) requires understanding AI's societal impacts, and ISO 31000:2018 mandates a thorough grasp of the organizational context. Meta-synthesis provides a robust evidence base for these standards by integrating qualitative data on stakeholder concerns and ethical issues, complementing traditional quantitative risk assessments and supporting the context establishment requirements of ISO/IEC 42001.

Enterprises can apply meta-synthesis for AI risk management in three steps. 1) Scope Definition: Clearly define the qualitative risk issue, e.g., "trust and bias concerns in generative AI for customer service." 2) Systematic Data Integration: Collect and screen relevant internal and external data, including academic papers, industry reports, and customer feedback. 3) Thematic Synthesis: Code and analyze all data to extract recurring themes (e.g., "inconsistent responses," "privacy fears"), translating them into specific risk scenarios for the risk register. A global financial firm used this method to analyze customer interviews on AI advisors, identifying trust gaps due to cultural differences. They adjusted their AI models, reducing related complaints by 15% and increasing service adoption, thereby meeting the risk treatment requirements of their ISO/IEC 42001-aligned AI Management System.

Taiwan enterprises face three main challenges. 1) Scarcity of Quality Data: A lack of localized, in-depth qualitative research on AI ethics. The solution is to establish internal data collection, such as regular stakeholder focus groups, and collaborate with academic institutions. 2) Methodological Expertise Gap: Corporate risk teams often lack qualitative research skills. This can be addressed through expert consulting (like Winners Consulting) and developing standardized operating procedures (SOPs). 3) Quantitative Bias in Management: Decision-makers may be skeptical of qualitative findings. To overcome this, link qualitative themes to key performance indicators (KPIs), such as connecting "low user trust" to "customer churn rate," and use visualizations like heat maps to present the severity of qualitative risks. A priority action is to run a small-scale pilot project on a specific AI application to demonstrate value within 90 days.

Winners Consulting specializes in meta-synthesis for Taiwan enterprises, delivering compliant management systems within 90 days. Free consultation: https://winners.com.tw/contact