Samsung AI vs. Non-AI Patent Correlation Matrix, 1982-2018

A 9.5 KB Excel file contains a correlation matrix derived from 10,695 patents filed by Samsung Electronics. This dataset supports a firm-level analysis comparing the pervasiveness and technological complementarity of AI versus non-AI patents. The research, authored by Sangrok Lee, uses patent data from 1982 to 2018 to empirically test AI's characteristics as a general-purpose technology.

Use Cases

• Analyzing the technological impact and generality of AI patents based on the patent-level metrics described.
• Studying systematic complementarity between technologies using the correlation matrix structure.
• Conducting robustness checks for firm-level innovation studies based on the alternative matching specifications mentioned.
• Investigating the GPT-like characteristics of AI through empirical patent data analysis.

Strengths

• Focuses on a specific, well-defined firm (Samsung Electronics) over a 36-year period (1982-2018).
• Analysis is based on a substantial sample of 10,695 patents.
• Results are described as robust across alternative matching specifications and variable operationalizations.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset's 9.5 KB size indicates it is a summary or derived matrix, not the raw patent records.

Provenance

Source

Patents filed by Samsung Electronics.

Collection Method

Firm-level patent analysis with matching of AI and non-AI patents.

Time Range

1982 to 2018

Freshness

Last updated 2026-05-05 17:32:12; freshness should be verified.