Miss something this week? Don’t panic. The CBC’s Marketplace roundup is here to help you catch up on the consumer and health news that actually mattered. This was a week of big warnings and hidden forces, the kind of week that makes you want to put your phone down and pick it back up slowly. It started with a disturbing conversation inside the artificial intelligence industry that spilled into public view. A researcher named Jacob Coxon announced on X that he had resigned from Anthropic, an AI company, because he was deeply worried about what he saw happening inside the companies building the technology. Coxon, who said he spent the past three years working at both Anthropic and OpenAI, accused the industry of “gambling with our lives.” His central message was blunt: the people building AI “earnestly believe that it could kill us all by the end of the decade.” That sounds like a doomsday headline, but Coxon insisted it wasn’t a marketing stunt. He said many executives and senior researchers phrase their concerns carefully in public to sound sensible, while privately they express genuine fear. The post got enormous attention, racking up more than 150 million views by Thursday. It also got a striking endorsement from inside the industry. Evan Hubinger, who leads alignment science at Anthropic, replied that the fear is real. “We really do earnestly believe AI could kill all humans!” he wrote. “I personally think it is >10% within the next decade.” Ethan Perez, a team lead at Anthropic, confirmed that Coxon had joined the company in May after working at OpenAI. The exchange opened a window into a debate that is usually kept behind closed doors: are the people building AI moving too fast, and are they being honest about the risks? These warnings are not coming from outsiders; they’re coming from people with their hands on the technology. It’s a question with no comfortable answer, and it’s only going to get louder.
The uncomfortable truth, according to a growing number of experts, is that AI development is racing forward faster than society’s ability to understand or control it. Coxon is not the first insider to raise alarms, and he probably won’t be the last. His resignation came at a time when AI companies are pouring billions into more powerful models, some of which are designed to improve themselves. The fear isn’t just that AI will make mistakes or take jobs; it’s that an advanced system could eventually do something catastrophic, and that the people in charge may not be able to stop it. That might sound like science fiction, but the people closest to the work are taking it seriously enough to resign, to post public warnings, and to argue with each other about what to do. Some of the replies to Coxon’s thread showed how complicated the situation is. Samuel Marks, a scalable oversight lead at Anthropic, responded in a personal capacity that developers keep pushing forward “due to a mixture of commercial incentives and a belief that they are in a race with other, less responsible AI developers that will abuse the technology or develop it less safely.” In other words, even the people who are worried feel trapped. They worry that if they pause, someone else won’t. The Anthropic Institute, the company’s research arm, said earlier this year that frontier AI development should be paused until regulations are in place, but acknowledged that would only work if every major AI lab agreed to stop at the same time. That kind of coordinated pause has never happened. Other industry experts have since chimed in on social media to say that these fears are common among developers, even if they don’t always say so publicly. The debate has drawn both skeptics who think the warnings are overblown and longtime risk-trackers who think the danger is real. So the industry keeps moving, and the warnings keep coming, with all the urgency and ambivalence you’d expect from people who see a potential cliff but can’t agree on how close the edge really is.
While the AI industry debated hypothetical risks to humanity, a much more concrete digital threat was being exposed in a joint investigation by CBC’s visual investigations team and Radio-Canada’s Décrypteurs. The target was a company in Vietnam called HTV Network, which presents itself as an “international media company” connecting the world with “authoritative journalism.” The reality, investigators found, is very different. HTV is actually an AI-slop empire, a factory of fake and sensational content designed to generate ad revenue by targeting vulnerable audiences around the world, including Canadians. At a corporate party in early 2026, the CEO, Ngo Anh Tan, stood on stage holding a sign that translated to “Record revenue.” And he had reason to celebrate: the company had grown to at least 100 employees, according to a Vietnamese job listing. But the celebration masks a murky operation. The investigation uncovered nearly 200 Facebook pages with a combined following of more than 55 million people, all sharing links to roughly 50 HTV-connected websites. Those sites run ads next to inflammatory, misleading, or outright false political and religious content. Some stories were bizarre clickbait, like a fake claim about Elon Musk banning Teslas in Canada. Others were darker, exploiting real tragedies like the Tumbler Ridge shooting to draw clicks and spread misinformation. “It’s quite staggering, the scope of what’s being done here,” said Fenwick McKelvey, a communications professor at Concordia University in Montreal. “[It’s] totally deceptive.” The deception extends even to the company’s name: HTV Network shares its name with a legitimate state-owned Vietnamese broadcaster, which makes it harder for platforms and users to trace the fake operation. When confronted, Ngo denied any connection to the Facebook pages or websites. Texting with CBC/Radio-Canada, he said he didn’t have any Facebook pages or websites at all. Asked why his phone number matched the one listed on the HTV website, he replied, “Maybe someone has taken my information.” In response to a written request for comment, he said, “I don’t understand what you mean.” Corporate records and domain registrations, however, matched his personal details. The investigation shows how the global misinformation economy works: shadowy networks, cheap AI tools, and social media algorithms combine to flood our feeds with junk, while the people responsible hide behind plausible deniability and reap the revenue.
In less existential but still stressful news, the Canada-U.S. trade war escalated again this week. The White House announced a new round of measures, including a ban on importing Canadian dairy, motorcycles, and some alcohol products, along with fresh 50 per cent tariffs on a slate of other goods. For Canadian business owners, especially those in the affected industries, this is another gut punch in a long, uncertain year. The scope of the trade war has grown, and regional disruptions could lead to job losses and the closure of small businesses. But if you’re looking at the big picture, economists say the actual damage to Canada’s economy may be smaller than the headlines suggest. Derek Holt, vice-president and head of capital markets economics at Scotiabank, crunched the numbers and found that the new tariffs apply to an estimated $3 billion worth of Canadian goods, including mattresses and some paper products. But at the same time, American officials dropped tariffs on other items, including toilet paper, cement, and sugar, which adds up to about $2 billion worth of goods. That makes the net impact marginal, especially when you consider that Canada exported more than $527 billion worth of goods to the U.S. in 2025. The bans on alcohol, dairy, and motorcycles are also negligible in the grand scheme, Holt said. Canada sends very little dairy and few motorcycles across the border. Alcohol exports are a bit higher, with $550 million worth sent to the U.S. last year, but the total value of Canadian items affected by the export bans is still only about $700 million. Holt called the U.S. actions “face-saving,” not substantive, and said that’s actually positive. The bigger economic risk, he argued, is the rising cost of oil, which soared past $100 a barrel for the first time since July due to escalating tensions in the Middle East. For everyday Canadians, the trade war remains a source of anxiety and frustration, and the unpredictability is itself a cost. The tariffs are also a reminder that trade policy can change with a single news cycle. Small businesses are making decisions with a moving target, and that’s hard to budget around. Even if the macro effect is muted, the emotional effect is not. But at least this round, the economic numbers suggest the sky isn’t falling, even if it feels that way.
A few other consumer stories also made headlines worth knowing. Grindr, the popular hookup app, agreed to pay a $48-million settlement over allegations that it violated U.K. privacy law by selling users’ sensitive information to third parties, including details about HIV status. The company disputes the allegations and says the settlement includes no findings or admission of liability. But the case is a reminder that the data we share on apps — even deeply personal health information — can be collected, packaged, and sold in ways we never agreed to. It’s an old internet lesson that keeps coming back: if you’re not paying for a service, you might be the product. Health privacy is an especially sensitive area, and this settlement shows that regulators are starting to pay attention. On a more educational note, a free course covering the basics of artificial intelligence will be available to all Canadians by the end of the year. K-12 teachers and post-secondary students will get access first on Sept. 21. The course won’t turn anyone into an AI researcher overnight, but it’s a practical step toward helping people understand a technology that is rapidly reshaping work, school, and daily life. For teachers, it’s a chance to get ahead of a tool their students are already using. And in the world of gadgets, Apple unveiled its new folding phone, the iPhone Duo, at an event that was also notable for being the first with new CEO John Ternus at the helm. The phone is sleek and futuristic, but it comes with a suitably futuristic price: at least $3,000. That’s a lot of money for a phone, and it raises the question of whether folding screens are a real revolution or just a luxury novelty. For most of us, the answer probably doesn’t matter, because the price tag alone puts it out of reach. But it’s a sign that the tech industry is still pushing boundaries, even as consumers tighten their wallets.
So what should you take away from this week’s news? Maybe a mix of vigilance and calm. The AI warnings are genuinely unsettling, especially when they come from people who actually build the technology. But it’s worth remembering that “more than 10 per cent chance” also means there’s a very good chance AI won’t destroy civilization in the next decade. The real day-to-day risks are more mundane: misinformation flooding your social media feeds, companies profiting from your personal data, trade policies disrupting your industry or your retirement savings, and new gadgets trying to convince you that you need to spend thousands of dollars. The best defenses are the same ones that have always worked: ask questions, check sources, protect your personal information, and don’t let alarming headlines make you panic. When you see a wild story or a shocking post, ask yourself who made it, why they made it, and what they stand to gain. That’s not cynicism; it’s self-defense in a digital world. The Marketplace roundup is here every Friday to help you cut through the noise and understand what actually matters. If you want it delivered to your inbox, the Marketplace newsletter is the easiest way to keep up. This week’s biggest stories may seem unrelated — a resigned AI researcher, a Vietnamese clickbait empire, another round of tariffs, a settlement over private data, a $3,000 phone — but they all connect to a broader theme: power is being concentrated in systems and companies that don’t always have your best interests at heart. The good news is that by paying attention, by staying skeptical, and by supporting journalism that digs into these issues, you’re already doing something about it. Don’t panic, but don’t look away either. That’s the balance. See you next Friday.