Peter Brantley
Director Online Strategy
University of California, Davis
The artificial intelligence (AI) services landscape is rapidly introducing new versions of old struggles over data and privacy. The release of OpenAI’s Model Context Protocol (MCP) as an open specification has led to a rush of new linkages between structured data sources and AI services, which can query, analyze, and link information resources in a common user-managed interface. MCP servers are beginning to populate cultural and research institutions, making available unique collections of content, their metadata, and applications available to the broader community of AI users. With the release of Apps in ChatGPT, enabling external applications to be called from within ChatGPT via MCP extensions and new APIs, large language models stand to aggregate and control a potentially vast array of user-generated intentional data. Almost 10 years since the release of Tim Berners Lee’s Social Linked Data (Solid) project, which aimed to empower users with the management of their own data, AI vendors stand on the threshold of creating new, greatly centralized data compilations that may offer some user management controls while raising impassable moats around data reuse. The briefing will give examples of MCP in action within the scholarly environment, while discussing the enduring issues of personal management of their information and its release. The presentation was co-produced with Gary Price (ARL Day in Review and infoDOCKET).
