Infrastructure is built before people fully understand what it will become.

Purpose: do we need AI data centres?

The debate is not just about chatbots and AI slop. It is about whether Alberta builds, governs, and benefits from the next layer of infrastructure.

Nobody needed today’s internet in 1994, when I built my first data centre, either β€” until they did.
Steve Keyser

β€œMy current data needs are already being met” is not how infrastructure works

Your current needs are being met because someone built fibre, towers, exchanges, cloud platforms, storage systems, and data centres before everyone personally understood how much they would rely on them.

AI infrastructure is at the same kind of inflection point. Some uses are shallow and annoying. Some uses will be serious enough that not having compute capacity will matter.

Real uses of AI beyond slop

🩺
MedicineDiagnosis, triage, imaging, treatment support
πŸ§ͺ
Drug discoveryFaster screening and simulation
β™Ώ
AccessibilitySpeech, vision, translation, support tools
🌾
AgricultureYield modelling, pests, water and fertilizer use
πŸ”₯
Wildfire modellingRisk prediction and faster response
☁️
WeatherForecasting and local prediction
⚑
Grid managementBalancing supply, demand, outages
πŸ€–
RoboticsSafer work and less drudgery
πŸ›‘οΈ
CybersecurityDetect threats and protect systems
🚚
LogisticsRoute planning and supply chains
πŸ“š
EducationPersonalized tools and tutoring
πŸ“
EngineeringDesign, simulation, testing
πŸ’»
SoftwareBetter tools and faster builds
🧭
Network operationsMonitoring, diagnostics, optimization
πŸ”¬
Scientific researchPatterns, models, and discovery

The junk people see first

  • Fake pictures and AI slop
  • Bad essays and copy-paste answers
  • Gimmicky features nobody asked for
  • Intrusive apps and shallow automation

Loud. Annoying. Easy to notice.

The risks are real

  • Surveillance and monitoring
  • Manipulation and misinformation
  • Privacy erosion
  • Labour disruption
  • Corporate control over essential tools

Serious. Complex. Needs governance.

Jobs: not simple

AI will take some jobs. It will change many jobs. It will also create jobs, but not always the same jobs, for the same people, in the same places, with the same skills.

That is why public benefit should include training, apprenticeships, procurement, education partnerships, and access to compute for Canadian research and businesses.

The answer is governance

Surveillance risk is a real objection. But the answer is privacy law, transparency, Canadian ownership, public-interest access, and accountability β€” not pretending AI disappears if Alberta refuses to host a building.

🧠

The real danger is hosting the infrastructure, carrying the impacts, and still watching someone else own the future.