Video of apparent use of AI by New Brunswick politicians goes viral – New Brunswick

A New Brunswick politician says he learned a lesson after using artificial intelligence to prepare a speech in the provincial legislature and admitting the video went viral.

Progressive Conservative Bill Oliver made a gaffe on June 9th during a speech he read from his notes, which has been shared on social media in recent days and garnered widespread attention.

This speech includes some strange references that turned out to be prompts from the AI ​​tools he was using. This includes a moment in his speech where he pauses and says, “This is a more natural, flowing version of that section, reading more like a speech in Congress than a series of short points.”

In a new statement released Wednesday, Mr. Oliver admitted that he had not carefully considered the text before giving the speech.

“When printing the final version of my speech, the AI ​​prompt was not removed, even though it was what I said and caused great concern to many people,” he said in a statement Wednesday.

But Oliver’s missteps have prompted warnings about the erosion of public trust caused by private uses of AI.

Mr. Oliver did not immediately notice the mistake, but when Congress recorded the entire incident, it became clear that no one else in Congress seemed to notice or react to his use of the AI ​​prompt.

“Although AI is becoming a more frequently used tool in many fields, it should not replace our ability to express our thoughts and beliefs,” Oliver said in a statement.

“The sentiment of my speech was certainly mine, and I learned an important lesson from this experience.”

The video spread around the world and made headlines on Al Jazeera, Euronews, India Today and Kenya’s Tuco News.

The song was not included in the viral clip, but Oliver later said, “Here’s a more developed, flowing version of this section.”

And on his third mention of AI, Oliver said, “I’m going to take this section with a grain of salt,” before interrupting and continuing to talk about the subject of affordability.

John Gillum, CEO of Collingwood, Ont.-based Originality AI, which uses software to detect AI for clients such as universities and news organizations, said he has seen many examples of “AI residue” in text.

But Oliver’s speech was unusual, he said.

“The confidence when AI residue is talked about in public is unique,” he said in an interview.

“I’ve never seen anything like this before. I certainly think it’s why it’s gotten so widespread.”

Gillum said his company analyzed about 8,000 English speeches delivered in the New Brunswick provincial legislature between 2022 and 2026.

It found that 3.3% of speeches could contain AI-generated text by 2022, up from 21% last year. So far in 2026, it has been found that nearly 23 percent of speeches are likely to contain passages from AI.

At least once, a politician asked a prepared question in parliament and received a prepared response, and originality analysis suggests both were likely written by AI.

“So it was a kind of political theater, with human actors reading the AI ​​story,” Gillum said.

Green Party leader Megan Mitton said Mr Oliver’s speech confirmed her suspicions that MPs had been reading AI-generated speeches throughout the first half of this year, which was a disservice to voters.

“We are MLAs. We were not elected to come here and read an AI-generated speech,” she said by phone Wednesday.

“We were elected not to be the voice of AI, but to be the voice of the people we represent.”

Mr Mitten said the Green Party caucus had rules banning speech writing by AI and supported a ban for all MPs.

The Liberal Party caucus said it had no policy on AI, but its representatives “have a responsibility to comply with existing obligations regarding ethics and professional conduct.”

Mitten said Congress needs “higher standards” and could show leadership by developing policy on the responsible use of artificial intelligence.

“If you look at the polls, we already have a crisis in our democracy, and I think using generative AI to replace our own thoughts will only deepen that crisis.”

Gillum said the use of AI models by politicians is problematic.

For example, Grok is controlled by former Trump administration official and billionaire Elon Musk, while models like OpenAI and ChatGPT may have different political leanings.

“Or is it China’s open model, controlled by political parties?” Gillum said.

“I think all of these things lead to the possibility of manipulation. So I think the best practice is to disclose the information.”

Sarah Laframboise, executive director of Evidence for Democracy, a nonprofit that tracks the use of science and technology in government, said it’s becoming common for politicians to use AI in speeches.

“The lack of human oversight is very problematic, especially when you’re talking about things that are on the legislative agenda or on the public record,” she said in an interview.

In March, Laframboise’s organization published a report on the use of AI in policymaking. Transparency was found to be “alarmingly low”.

He said the federal government is doing better than local governments, and said the online registry shows the full range of AI usage in the public sector, including government vendors.

But outside of campaigning, Laframboise says there are few AI guardrails for politicians.

He added that AI policy could be added to the code of conduct for federal lawmakers, but enforcement would be a big issue.

For example, some cell phone cameras use AI to process images, so politicians may not even be aware that they are using that technology.

“I don’t know if we have a clear understanding of how to regulate that yet.”