Mandatory ambiguity: a hybrid human–AI analysis of impact reporting by benefit corporations

Governments worldwide have increasingly established new legal forms intended to empower and govern hybrid organizations’ dual pursuit of economic and social value. Mandatory impact reporting is a centerpiece of these legal innovations, yet little is known about how hybrid organizations navigate the often-ambiguous reporting requirements these laws stipulate. Drawing on a unique dataset of all mandatory impact reports filed from 2016 to 2024 by the population of benefit corporations incorporated in Minnesota, we introduce a novel, human-derived scale of benefit report quality, encompassing five dimensions. Developing and applying a rubric-guided LLM scoring protocol to rate each report against the scale, we find that third-party certification by B Lab increases both the length and quality of benefit reports, thereby revealing a nuanced interdependence between state and private regulation. Our study highlights the limits and possibilities of mandatory disclosure regimes within hybrid legal forms while demonstrating the potential of rubric-guided LLMs to enhance content analysis.

Subjects

B Corporations

Benefit corporations

Certification

Hybrid organizations

Impact reporting

Public policy

DDC Class

320: Political Science

650: Management, Public Relations

Options

Mandatory ambiguity: a hybrid human–AI analysis of impact reporting by benefit corporations