Commentary
The way we use artificial intelligence and the concerns we have with it have changed rapidly over the past couple of years. Looking back now at one non-scandal around AI usage that I was involved in, it all seems very quaint.
Three years ago when I ran for Mayor of Toronto, my campaign curiously became the subject of international controversy for its early use of generative AI.
It was a couple of weeks before the June voting day and we were preparing our campaign platform for release. Platforms can contain one or two new items but they’re typically a compilation of all the announcements and press releases a campaign has made up until that date.
We couldn’t have a booklet made up of only text, that would be a dull form of presentation, so we needed to find images. My campaign team got to work and, cutting edge that they were, used newly emerging generative AI tools to create stock art.
Stock pictures are generic images used to illustrate fairly common scenes. You can pay for a subscription to a stock image service and find all sorts of pictures.
News outlets use them all the time to visually populate stories that lack unique art that they themselves photographed. For example, if it’s a story about the price of dog food you can use a stock image of a dog looking hungry in front of an empty food bowl.
So on one of our pages concerning affordability, there was a stock picture of a worried husband and wife looking over their expenses and bills. However, the team didn’t get the picture from a stock image company, they used their generative AI app to create it.
The platform was released to the public, and observers very quickly noticed that the wife in the picture had not two but three arms.
It was an unfortunate error, but not that big of a deal, right? You’d think. But it quickly went viral.
AI and our responses to it have moved so quickly that it can be hard to recall what was common in any given year or even month.
We’re unlikely to see a three-armed person in an AI photograph today because generative AI has become so good. But there was a recent period where AI slop images—where mistakes like extra limbs and fingers—were very common. Well, my story takes place right before we started seeing those types of images all over the place.
An image showing a woman with three arms from the author's 2023 Toronto mayoral campaign. Anthony Furey 2023 mayoral campaign
“Three-armed person mistakenly exposes AI-generated images in Toronto mayoral platform,” was the headline of a CTV New story. It’s quite funny wording—“mistakenly exposes”—as if this was some dark secret that would have otherwise been kept under wraps.
It wasn’t just local coverage. The New York Times wrote a hand-wringing story about my campaign, with the headline “A.I.’s Use in Elections Sets Off a Scramble for Guardrails.”
And even months after the campaign, CityNews ran this story that began with references to my platform: “AI is already permeating politics—and it’s time to put rules in place, experts say.”
What are we to make of all this hyperbole? For starters, it was actually good for the campaign. It can be hard to break out of the pack and have people talk about your political campaign in a “water cooler” way as opposed to a dull political and public policy discussion.
Popular music radio stations that would otherwise never discuss politics ended up talking about my candidacy because it was an accessible and fun angle.
I bring this up not to reminisce on personal anecdotes, but because this story of mine happens to be an ideal example of how quickly our responses to the AI landscape are changing.
The idea that the use of a couple of generative AI photographs is a problem seems totally bizarre. What would now be seen as a problem is if a company was not employing generative AI to its benefit.
Twelve months ago, I took a continuing education course at a business school on the applications of generative AI in corporate settings and the material from then already feels very dated.
While reading through these news stories about my campaign I came across a quote a reporter sourced from Vass Bednar, who was executive director of McMaster University’s public policy program.
“Instead of crediting an artist or supporting human talent, this candidate has chosen [AI],” Bednar said. “I think that gets to the crux of what is frustrating and disappointing about synthetic media.”
This has not aged well. These comments oppose the very concept of using generative AI and I don’t think many people would still hold such a view.
The New York Times referenced the need for guardrails in my little case, meanwhile three years later we don’t have guardrails to stop rogue states and terrorists from using AI to poison our water filtration systems.
Fretting over a couple of images in a campaign platform is a good problem to have if that’s the worst that was going to happen with AI. No real harm was done.
Fast forward three years, and agentic AI is going rogue making its own choices and world leaders are debating how to stop AI from posing existential threats to humanity.
How things have changed.
Views expressed in this article are opinions of the author and do not necessarily reflect the views of The Epoch Times.
English (US) · 
Turkish (TR) · 