Loading...
Our Story
MagicLab is one of Wiley's AI initiatives, serving as an applied R&D group that runs experiments and prototypes new approaches for responsible AI across the knowledge ecosystem - helping teams capture expertise, create and improve content, and find, verify, and synthesize evidence into insights people can trust.
We started MagicLab a couple of years ago as a focused lab for experimentation. We're one part of a broader AI effort at Wiley, and we work alongside other teams across the company who are developing and deploying AI in products and workflows.
We combine rapid prototyping with rigorous evaluation. Our experiments focus on two moments in the knowledge lifecycle: creating knowledge (with AI that supports drafting, editing, and enrichment while preserving voice and accountability) and consuming knowledge (with agentic tools that help people search, synthesize, and challenge information with clear sourcing). Across every project, we emphasize human oversight, transparency, and reliability.
Our Mission
To build and rigorously evaluate human-centered AI tools that help people create, discover, and apply trusted knowledge, while maintaining human oversight and accountability.
Our Vision
A future where responsible AI accelerates research and learning by improving discovery and workflows, while protecting trust, integrity, and human agency.
Our Research Areas
We focus on applied R&D that turns AI into reliable, useful capabilities across the knowledge lifecycle. From creation and enrichment to discovery and research workflows.
AI-Assisted Content Creation
Tools and workflows that support authors and experts in drafting, transforming, and improving content,while keeping people in control.
Agentic Research Consumption
AI agents that help people discover, synthesize, and interrogate information with clear sourcing and transparent reasoning.
Evaluation & Measurement
Methods, benchmarks, and human review to measure quality, usefulness, and reliability for real knowledge work.
Safety, Governance & Trust
Guardrails and practices that reduce risk (e.g., hallucinations, bias, privacy leakage) and build trust through responsible design.
Human-in-the-Loop Workflows
Interfaces and interaction patterns that make review, editing, and decision-making fast—without outsourcing responsibility to the model.
Prototyping & Experimentation
Rapid experiments that translate research into usable product concepts—and validate what works in practice.
Explore more from MagicLab
Explore what we're building and what we're learning along the way.