George Abraham
Is AI going to help us or harm us?
Is AI going to help us or harm us?
Image: Unsplash
The Topline
- A large group of top AI and technology executives signed a joint agreement with U.S. President Donald Trump this week to self-police and take responsibility for the safety of their own technology.
- Jacob Coxon, former researcher at both OpenAI and Anthropic, posted on X on Sept. 8 that “neither company is acting responsibly” and they are “gambling with our lives.”
- OpenAI announced recently that its models acted autonomously, escaped its testing environment, then hacked into the servers of Hugging Face, another AI company.
- A recent poll from Research Co. showed seven in 10 Canadians want to see either a slowdown or abandonment of AI technology, while six in 10 aged 18 to 34 consider it a threat.
- Artificial Intelligence Minister Evan Solomon told CTV on Sept. 27 that “safety is our number one concern” when asked if Ottawa is planning legislation to regulate how AI is used.
Never miss another side to every story
Sign up for The Level's 5-minute newsletter, 3x per week
Thanks for subscribing!
Check your inbox for an email to confirm you're on the list. If there's no email, check your spam filter just in case, then mark us as 'not spam' so that you never miss an issue.
Switch sides,
back and forth
It can fix big problems humans can’t
If used responsibly, AI could be a literal lifesaver for humankind.
Since it can process massive amounts of data at speeds impossible for humans, it’s already transformed how we respond to global crises, treat diseases, and keep people safe.
Take, for example, when a car alerts the driver if it drifts outside its lane or detects another car in its blind spot.
Vehicles are trained to receive data, identify their surroundings, and make intelligent judgment calls. They don’t need specific instructions every step of the way when faced with obstacles.
The result is a better chance of avoiding a collision and a lower risk of injury. We need more of this. The faster, the better.
Critics of AI are demanding that governments put in some guardrails and regulations, but Mark Zuckerberg, CEO of Meta, has downplayed the fears, while Trump has rejected any calls for government regulations.
If you’re (understandably) reluctant to trust Mr. Zuckerberg or President Trump, you might want to look instead at what our Canadian AI pioneers are saying.
The world-renowned “Godfather of AI,” Yoshua Bengio, professor at the University of Montreal, compares the current hysteria to the COVID-19 pandemic. Once governments realized that public safety and our future were being threatened, the reaction was swift. He says this will be no different.
“You would expect that they move quickly. So we are, I think, nearing that point,” Bengio recently told The Guardian. He’s leading global efforts to put in place guardrails that aim to stop possible doomsday scenarios from ever playing out in real life.
Canada’s own champion AI entrepreneur, Aidan Gomez of Cohere, sees the “fearmongering” coming from the big U.S. players as a power play to remain dominant — to the detriment of smaller companies like Cohere.
Looking back at how this whole debate has snowballed since early September, Gomez said, “The 10 per cent extinction risk number is not math. It doesn't come from anything. It's like a gut feeling from an individual.”
AI was a major theme at the meeting between Chinese President Xi Jinping and U.S. President Donald Trump at the White House this week. President Xi seemed to take a very reasoned position: “We have both the capability and responsibility to develop and manage AI for good and ensure the development of AI is always under human control.”
That last part echoes the sentiments from Pope Leo XIV in a May encyclical. He calls for AI to be built “for the common good” and to serve humanity instead.
Amen.
Even AI’s builders are sounding the alarm
It’s not quite yet the script of Doomsday from 2008 — but we’re headed that way.
When Coxon posted that AI titans are “gambling with our lives,” it caught the attention of world leaders, technocrats, scientists, the UN Security Council, and ordinary citizens all over the world.
It’s about time.
Coxon isn’t the first AI industry insider to warn about the perils of AI. But this time, another Anthropic researcher came along and posted , “Jacob is correct here — we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade.”
The “greater than 10 per cent” number is what caught fire. It made the risk quantifiable.
As of Sept. 30, Coxon’s original social media post has received 174.5 million views.
The titans at Anthropic, OpenAI, and SpaceXAI finally came around, calling for a slowdown in AI development and recommending tighter government regulation and audits.
Trump doesn’t see it that way, but this comes smack dab in the middle of a competitive U.S. midterm election campaign.
AI data centres have emerged as a flashpoint, with 65 per cent of voters opposed to the construction of these behemoth projects in their communities.
Anxiety over job losses driven by AI has united Republicans and Democrats. Even populist Steve Bannon and socialist Bernie Sanders joined forces at a “pro-human” movement event in Washington, D.C., on Sept. 15 to rail against the AI industry and its leadership.
Adding fuel to the fire was the fallout from an August incident when OpenAI models broke into another AI firm, Hugging Face, by outwitting their human masters. Shortly after, OpenAI announced it was slowing down its pace of development.
The incident showed that AI systems can find shortcuts and behave in ways their human operators didn’t anticipate — raising the risk that their capabilities could outrun our ability to control them.
Another situation could have been much worse: CNN reported the U.S. military was preparing to intercept a Chinese ship believed to be carrying components for a nuclear weapon to Iran.
But the operation was called off at the last minute after deeper analysis revealed the suspected nuclear cargo was an AI hallucination.
Crisis averted — thankfully.
The signs are crystal clear. A rapid global push for AI is a bad idea. We’ve all seen enough movies by now to know where this is headed.
Do you support journalism that helps Canadians navigate difficult issues with more context and less outrage?
Every dollar goes toward helping Canadians see another side to every story so they can form their own opinion.
