For all the talk about AI as something the world has never seen before, many of the problems it's running into are ones other industries have faced before. You just have to know where to look.

A few weeks ago, Chris Hogg was at Upper Bound, an annual AI conference hosted by Alberta Machine Intelligence Institute (Amii) in Edmonton. During the event, four oil and gas veterans took the stage and politely told the AI crowd to look at their scars first. You can read more from Chris about it here.

Sixty years of pipelines and power plants taught them what goes wrong when you build something big and forget to ask the people living next to it. 

 Now they're watching the data centre boom make the same mistakes. 

Marla Orenstein of the Energy Futures Lab said if you went hunting for the industry that most resembles where AI is headed, you'd point to energy. The gold rush, the money showing up before the rules, the bills nobody totalled until years later.

Canadians are already uneasy about it. An Abacus Data poll in March found just 16% would welcome a data centre in their community, with another 39% on the fence. 

Rocky View County, northeast of Calgary, voted down a 448-hectare campus 6 to 1, then wrote its own bylaw on setbacks, noise, and public engagement so the next developer knows the rules before they show up.

In energy, as in data centres, the people who carry a project's costs without sharing the rewards are the ones who can stop it. Peter Tertzakian, a Canadian energy economist and co-host of the ARC Energy Ideas podcast, put it plainly at the panel.

"The person who doesn't get a cheque has the ability to disrupt the whole project," he said.

Kevin O'Leary, the self-styled Mr. Wonderful, is being sued over his Wonder Valley project near Grande Prairie by a First Nation that says it was never consulted, which doesn’t make it seem like a wonderful neighbour. 

A couple of hundred kilometres north, Woodland Cree First Nation owns 51% of a data centre on its own land, with the revenue going to housing, schools, and elder care. When the people next to the project share in it, it becomes something worth supporting rather than something to fight.

Strip it down, and the problem is that the technology and the money behind it are moving faster than the people they affect.

The same thing happens inside companies. Roll out an agentic workflow or restructure a team around AI, and the person who can stall it is down the hall, not at a zoning hearing. 

The energy sector paid full price to learn to ask first. For everyone else, it's on sale.

Some light reading for when the zoning hearing runs long

Canada's new AI strategy is ambitious and the hard part starts now. My colleague Jennifer Friesen covered the launch of the $2.3 billion federal strategy, which aims to increase Canadian AI adoption from 12% to 60% by 2034. It funds compute, institutes, and capital, and goes quiet on who makes any of it work inside an organization.

When an agent owns the work, who is responsible? Agents are now running whole workflows, and the question of who answers when one carries the work falls to the CIO, who gets the responsibility without the authority to match.

When your data has a lifeguard problem. The City of Kelowna asked its chatbot which beaches had lifeguards, and it confidently named ones that had none, off a forgotten PDF. In another piece by my colleague Jennifer Kervin, a clean look at why AI sits on top of unglamorous data work.

What happens when production planning gets too big for spreadsheets. A Winnipeg brewery ran its entire production schedule out of one founder's head and a spreadsheet, until it seized up. If your operation depends on a file called Final_FINAL_v7, this one's a mirror.

Snowflake's new AI tools need data many companies lack. I was in San Francisco for Snowflake Summit, where the theme was that the platform holding your data should run the AI that acts on it. In the company's own words, there's no AI strategy without a data strategy, which is a polite way of saying the homework comes first.

The most hireable IT skill in 2026 is being human. Nearly nine in 10 IT leaders expect to restructure this year, and the capabilities rising to the top are the ones AI can't copy, like judgment, communication, and reading a room.

Why the TTC opened its doors to an ecosystem to modernize. A 106-year-old transit agency that still keeps a blacksmith on staff is testing AI, drones, and digital twins by turning its own system into a sandbox for startups and researchers. 

ScaleUP Awards recognize Western Canadian companies. Eight winners across the West, $640 million in combined revenue and more than 1,800 jobs, from an Indigenous food manufacturer in Alberta to a Vancouver ventilation company whose systems run in the Empire State Building and a wing of the Pentagon.

It's easy to keep your head down in your own industry and assume your problems are brand new and uniquely yours. They almost never are. 

The energy sector has already lived through what happens when you skip the awkward conversations, and a lot of what AI is learning the expensive way, another industry figured out years ago.

Ferris Bueller had it right. Life moves pretty fast, and if you don't stop and look around once in a while you could miss it. 

The lesson you need before your next rollout is probably already out there, in another industry or another part of the country, waiting for you to go find it.

— David

Most AI tools are a trade — your data for intelligence. Norton Neo breaks that deal. Powerful built-in AI, anti-fingerprinting, VPN, and ad blocking come standard. No setup. No add-ons. No compromises. Search, summarize, and write with AI that works inside your browser and stays there.

Browse Smarter, Stay Private