PUBLIC INPUTAI should say when it’s AI.
Canada · Lead story
Canada wants clearer AI disclosures. Here is what useful transparency looks like.
Labels, incident records and agent logs could change how Canadian organizations design AI products.
Today’s briefing
QUICK TAKES
Sovereign AIWhy compute is now industrial policyRAGThe work that happens before the model writesEU AI ActWhat Canadian exporters need to knowTHE HOME DESK
All Canada coverage →Canada is building its AI playbook in public.
CanadaCanada watch
Canada wants clearer AI disclosures. Here is what useful transparency looks like.
Labels, incident records and agent logs could change how Canadian organizations design AI products.
CanadaCanada watch
Sovereign AI compute is becoming Canadian industrial policy.
Who gets access to chips, power and data-centre capacity may shape the next generation of Canadian companies.
CanadaCanada watch
What Canada’s AI Safety Institute can actually do.
Testing, measurement and international coordination matter only if findings reach deployers and the public.
MODELS & PRODUCTS
Model tracker →What changed—and what’s just a launch-day claim.
ModelsModel briefing
The real test for a frontier model is whether it can finish the work.
Long tasks, tool use and reliable revision matter more than a single impressive answer.
ModelsModel briefing
Why the workhorse model may matter more than the flagship.
ModelsModel briefing
Multimodal AI is becoming one model, not five separate tools.
ModelsModel briefing
Reasoning models trade speed for a better chance at solving the whole problem.
DEEP READS
Browse all →Follow the evidence.
BusinessMarket signal
The race to connect AI agents is becoming a platform business.
SDKs, tool standards and permissions determine whether an agent can safely reach the systems where work happens.
BusinessMarket signal
Inference economics will decide which AI products survive.
A popular feature can still be a bad business if every answer consumes more value than it creates.
BusinessMarket signal
AI chips are now a strategy question for every serious deployer.
Hardware availability, memory and networking shape cost and model choice long before users see a feature.
BusinessMarket signal
AI data centres are turning electricity into a board-level constraint.
Power availability, grid timing and community impact now influence where compute can grow.
Primary sources first. Clear labels. Corrections in public.
AI moves fast; accuracy still matters. Our news analysis links to the original announcement and separates reported fact, company claim and AI New analysis.
Read our standards →