Topic Hub
Evidence, Research & AI
What do we actually know, and how far can the evidence travel?
The public discussion about AI often moves faster than the evidence. Immediate performance becomes a claim about durable capability. Association becomes causation. One occupation becomes a statement about professional work generally.
This hub is where those boundaries stay visible.
01 · Start here
What do we actually know?
What Do We Actually Know About AI and Professional Capability?
A research-led review of performance, learning, calibration, human–AI collaboration and the limits of current evidence.
02 · Core arguments
The core arguments.
Evidence, Inference and Uncertainty
A practical convention for separating what research observed, what we infer from it and what remains unresolved.
Cognitive Offloading, Fluency and the Judgment Gap
What the evidence suggests about external assistance, confidence, calibration and the distance between output quality and evaluative capability.
When AI Synthesises Too Early
Why research synthesis should preserve disagreement, provenance and epistemic sequencing.
Verification Is Becoming Part of Professional Work
How evidence, standards and consequence shape the depth of scrutiny required before an output deserves trust.
03 · Questions people are asking
Questions people are asking.
- Does AI actually make professionals more capable?
- Is critical thinking declining, or merely changing?
- What can a productivity study legitimately tell us about learning?
- How do we read AI claims without becoming paralysed by uncertainty?
- When does a plausible synthesis outrun the underlying evidence?
04 · Research and scholarship
The wider research agenda.
The current research programme examines capability, judgment, professional formation and human–AI workflow governance while stating its developmental stage explicitly.
05 · Continue exploring