Research & Intelligence

Evidence before commentary.

Scottish AI briefings, reality checks, case studies and primary-source research. Facts are separated from inference, and limitations are published.

Programmes

Research streams

State of AI in Scotland

Quarterly report combining adoption, policy delivery, infrastructure, business and public-sector evidence into one readable picture.

Quarterly

AI Reality Checks

Repeatable tests of products and claims, including failures and limitations. We test what vendors and headlines claim.

Ongoing

Scottish Case Studies

Measured before/after results from real workflows, with the method and caveats published alongside.

Ongoing
Methodology

Public-sector AI register

A reproducible method for recording public-sector AI use cases, procurement and evaluation. Built so others can audit and extend it.

Record

Capture the use case, owner, data involved and procurement route from public sources.

Verify

Link every entry to a primary source and the date it was checked.

Evaluate

Note what is measured, what is claimed and what remains unverified.

Publish

Release the register and its method so others can reproduce and extend it.

Full method and source grading on the sources & methodology page.

Editorial standard

What you can expect

Primary sources first

Government, official and primary documents before commentary.

Exact dates

Every figure carries the date it was checked.

Facts vs inference

Observation is separated from interpretation.

Corrections versioned

Material errors are corrected publicly and the change recorded.