The artificial intelligence cooperative: READ-COOP, Transkribus, and the benefits of shared community infrastructure for automated text recognition

This research article, published by Open Research Europe, discusses the formation and activities of READ-COOP, a cooperative dedicated to advancing Artificial Intelligence (AI) and Machine Learning (ML) technologies, particularly in the realm of Automated Text Recognition (ATR). The authors include Melissa Terras, Bettina Anzinger, and Paul Gooding, along with several other experts from various European institutions, underscoring a collaborative effort in promoting responsible AI frameworks.

READ-COOP serves as a platform cooperative that has developed its own AI tools, specifically within the cultural heritage sector. The cooperative currently boasts 227 members from 30 different countries and has registered 235,000 user accounts. Its primary tool, Transkribus, has processed 90 million digital images of historical texts, showcasing its significant impact in the field of digitization and historical document transcription.

The cooperative model facilitates democratic decision-making and emphasizes stakeholder involvement, which has proven beneficial for both innovation and community engagement. Members express high levels of satisfaction regarding the governance of READ-COOP, indicating that cooperative principles contribute to sustainable growth and effective management of AI infrastructures.

The study employs qualitative methods, including member questionnaires and reflection-in-action, to assess the cooperative’s development from a European Commission-funded project to an independent entity. It highlights the cooperative's structure, membership dynamics, and operational efficacy from 2019 to 2024.

READ-COOP's cooperative framework promotes accountable AI development by prioritizing member needs and community involvement. The cooperative has won the European Union's Horizon Impact Award and demonstrates that a cooperative business model can effectively support innovation in AI and ML, particularly in cultural heritage contexts.

The cooperative's tools have been credited in numerous academic publications, reflecting their widespread acceptance and usage in academic and cultural institutions. READ-COOP’s approach effectively demonstrates how cooperative governance can enhance technological innovation and community engagement while ensuring the sustainability of digital infrastructures.

The cooperative’s governance aligns with ethical principles in AI, establishing a framework that emphasizes transparency, equity, and community concern. By engaging members in decision-making processes, READ-COOP is positioned as a model for the responsible deployment and governance of AI technologies in the cultural sector.

The findings advocate for the broader application of cooperative models in AI and ML technologies, especially those initially funded through public grants. The article suggests that cooperative frameworks could provide a sustainable transition from public development to long-term community ownership, ultimately fostering ethical and responsible practices in technology governance.