This is Vanese Ferguson with Coffee Talk. The NDP is not wrong when it suggests decision making around AI data centres have been unilateral in nature. The sudden uptick in A-I -everything, or what seems sudden to me, got me thinking about how long ago did we really started talking about AI literally in everything, all the time. In terms of our news stories, it appears to date back to 2023-ish when we ran a story about AI being used in cardiovascular care in Saskatchewan. Subsequent earlier stories were equally benign like the City of Saskatoon using AI in a residential energy map to indicate home energy consumption estimates or using artificial intelligence to develop vaccines for future pathogens at VIDO on the USask campus. But that’s the point, isn’t it? AI was being used as a support tool not as the main prong of everything we do. I read a story yesterday that frankly I found chilling. Researchers took some of the leading AI agents – that’s how they are referred to, like Grok, Gemini, Open AI, as agents, and put them in virtual real-world societies to watch them interact. But many of them actually developed their own shorthand speech that had meaning to the AI agents but the humans monitoring the experiment couldn’t decipher what that meaning was. Which brings to light the bigger concern of maintaining control over AI usage. And highlights, in my mind, the lack of input and control Saskatchewan residents have had when it comes to developing AI data centres that at first blush may mean jobs and economic spin off but in the long run could have significant implications. If we were to have a referendum in Saskatchewan today – that would be my topic of choice.