One question per article. Answered properly.
Long reads from our editorial team on how AI systems work, where they fail and what that means for you. Free to read, no sign-up.
All articles
- AI apps and your privacy in Canada: what data they collect and what rights you have What AI apps may collect when you chat, upload or speak, what PIPEDA says about consent and purpose, and how to ask an organization for your information.
- How banks use AI to flag unusual payments When your bank asks you to confirm a purchase, a model has probably compared it with your usual pattern. The concepts behind those checks, without jargon.
- Why AI chat assistants sometimes make things up Chat assistants predict likely words, not checked facts. That is why they sometimes invent a quote or a source, and how you can catch it.
- Training data explained: why what a model learns from matters A model can only be as good as its examples. Where training data comes from, how labels are made and what gaps do to the results.
- How image recognition works: from pixels to patterns A photo is a grid of numbers. How a model turns those numbers into edges, shapes and finally a label, and why a sticker can fool it.
Try this: pick the article closest to an app you used today and read only its first section. Then decide if you want the rest.
How every article is built
- One clear questionThe title is the question. The article answers it and stops.
- A picture of the ideaA diagram, a step list or a set of definitions you can come back to.
- The limitsWhere the technology fails, who it affects and what is still uncertain.
- A Try this taskSomething small to do with your own phone or browser after reading.