Commissioner of Canada Elections Will 'Explore the Use' of AI

The Office of the Commissioner of Canada Elections (OCCE) has revealed in its annual report that it will "explore the use" AI and emerging technologies to see how they will shape the government body's approach for the next year. From a report: Commissioner Caroline Simard's office didn't outline ways it might adopt AI. In its outlook, the OCCE expected to use funding announced in January 2025 to secure the tools needed for addressing the "challenges of today's electoral environment." This included staffing roles dictated by its new structure and reflected "ongoing modernization efforts," but no further details.

The Commissioner is an independent officer who ensures the government, political parties, and others honour both the Canada Elections Act and Referendum Act. This includes core aspects like financing, nominations, campaigning, and advertising. More recently, the OCCE has been addressing rising issues with AI, including election disinformation facilitated by bots, AI-generated images, and deepfakes (AI-generated videos that resemble real people in false scenarios).

So this is how democracy dies

[rsilvergun]

It's not thunderous applause it's not cool it's not fun it's not exciting. It dies in committee hearings often at the county level. It dies from shitty little assholes taking over what should be nonpartisan administrative roles and putting sycophants in charge top to bottom.

People expect the death of democracy and the coming of fascism to be something they can see and feel and something that is on par with storming the beaches of Normandy because that's what television told them and if you grew up with

Re:So this is how democracy dies

[Panspechi]

I'm not sure why this rant is a response to the article, but the Office of the Commissioner of Canada Elections is simply following a broader Government of Canada policy to see how LLMs can impact the work of public servants and/or Canadians in general. We have few laws against deepfakes and our government is looking at what might breach them and how to adjust the current laws to improve the integrity of elections. I'm a federal employee in Canada and our intranet is a big mess. I'm happy to be part of a pilot project for an internal use of a LLM chat agent to help employees find the info/forms/software they need that the department already possess. All data is stored within our department's servers and queries retention is 14 days so I can return to them, then it is deleted everywhere. I'm pretty sure our government is not planning to release deepfakes of Trump being arrested by the RCMP...

Um.. that was my entire point

[rsilvergun]

So AI in elections is a bad thing. That's not something I'm going to debate. LLMs are far too complex to allow for any kind of transparency and elections live and die by transparency. So llms do not belong in elections. Full stop.

So you have a government rule instructing everyone to use AI which means llms and machine learning too but same difference. A rule that you can probably Trace back to billionaire owned companies lobbying for it. But it's a rule that's going to be enforced and built up on the gr

And the Elon wil surely go to jail because

[Anonymous Coward]

He is a proven saluting nazi fascist. He manipulated the election with the help of AI and gerrymandering maps preventing 7 millions democrat voters to vote. Somebody ran the numbers to figure out how much voters suppression there was. In 2024 7 million democrat people tried to vote and couldn't.

The Democrats could run Jesus fucking Christ and it wouldn't do any good if the Republicans just stop 7 million people from voting.. it literally doesn't matter how good a candidate they run. It doesn't matter what t

Re:

[rsilvergun]

My friends, I have a request. That request is, for all that’s Holy, stop calling things “clickbait.” It’s an old and busted term that has no place in the media landscape of 2023.

It’s clickbait all the way down For starters, calling something clickbait implies that trying to draw attention to content is somehow inherently wrong.

Literally every article and blog post on any publication or (these days) corporate sponsored site is clickbait designed to get people to find and r

Re:

[Quasar1999]

Logic is a systematic method of coming to the wrong conclusion with confidence.

Are we really ready....

[Roger W Moore]

...to welcome our AI-elected overlords?

I'm having a lot of trouble seeing where AI can be useful in an electoral system. The key feature of an election is that an average member of the public must be able to know that it was fair so they can accept the result. This means it not only has to be fair but also seen to be fair. AI algorithms are usually completely opaque even to experts in the field and their fairness depends entirely on their training which again is not something a member of the public can readily determine was fair.

Re:

[ClickOnThis]

Unlike many people commenting on this story, I actually read TFA. I didn't see anything about using AI to change the way elections are run, specifically counting ballots cast by human beings, one each.

What I did see is a government agency exploring ways that AI could help run elections (such as making government documents easier to find) and by examining how malicious AI by third parties could threaten the integrity of an election (such as with disinformation and deep fakes.) Nothing has been decided yet, a

NOOOOOOOO

[Big Bipper]

There's too much lying going on in our elections already. Bring back paper ballots and scrutineers. It won't hurt us to wait till the next morning to find out who really won.