What is MagicLab?

As AI agents enter the (mostly human) scientific research and publishing network, existing actors in that network face substantial disruption. This ranges from the obvious -- how will we distinguish good science from junk? -- to more future-oriented concerns (if AI agents "read" papers written by other AI agents, why have papers?). Magic Lab is a space for the research community to engage with these disruptors outside the bounds of Wiley's journal publishing products.

Our Goals

While some AI tools or techniques facilitate feature improvements for a current Wiley product, others challenge either the structure of the market for that product or the role of human actors in that market. Magic Lab exists to investigate ways of addressing those challenges that might make the network stronger, increasing circulation among its actors, without cutting humans out of the heart of knowledge production.

The Lab is not a proof-of-concept factory, and it's not a marketing site for feature releases: it's a place where Wiley's AI developers can share problems and learnings (the "insights" part of this site) as well as invite participation in "experiments" with our technology. We commit to sharing the results of these experiments publicly and transparently with an aim toward building consensus on how we -- Wiley, researchers, readers, tenure bodies, grantmakers -- should permit or constrain AI participation in the research network we've built together over hundreds of years.

An Actor-Network for Research Publishing

(insert some material from Andrew's white paper outlining an ANT analysis for authoring, consuming research content)

Authoring, Grantmaking, and Tenure

Diagrams, etc.

Research and Manuscript Preparation

Diagrams, etc.

Major Research Agenda

An overview of stuff we want to build, experiments we want to run.