Table 1 Workflow of the analysis of GAI/LLM industrial guidelines and policy statements.

Data preparation

Focused on selecting 160 top-tier companies across 14 industrial sectors globally.

Ensured a diverse representation of countries and regions.

Data preprocessing

Excluded companies without official guidelines or policy statements.

Supplemented the dataset with interviews or company statements when official policies were absent

Systematic review

Categorized and thoroughly analyzed guidelines to evaluate the scope.

Assessed ethical considerations and regional/sectoral implications for using GAIs and LLMs.

Qualitative analysis

Identified major themes, concerns, and divergences in guidelines across sectors

Emphasized region-specific and industry-specific challenges.

Quantitative exploration

Tokenized the text using NLTK. Used sent tokenize for sentence splitting. Used word tokenize forward-level tokens.

Removed stop words to focus on substantive content.

Text mining-based analysis

Built a TF–IDF model for the corpus, estimated term importance across industry guidelines.

Compared term weights across sectors, surfaced sector-specific keywords for downstream analysis.

Clustering and pattern finding

Applied K-Means clustering to identify and visualize common patterns and themes across guidelines

Used clustering visualizations to highlight cross-sector differences.