Canada’s Safe Social Media Act, Just the Facts

**Taylor Owen** (0:04)
I open up my XFeed right now, and there is not a reasonable debate going on about the nature of this bill.
I am seeing the loudest, most negative voices or the most boosterish voices on either side. I am not seeing a discussion about what's actually in it. And the media isn't particularly good at having that discussion either.
We are heading into a world of rapid technological change driven by AI. And the public, I don't think, has the means or the media it needs to help them understand what we are all going through collectively. And that's a problem in our daily lives and in our businesses. But I think it's also a problem for policy. Because policy is big and complicated and boring. And none of those things work on social media. What works on social media is anger and outrage and hatred. And that is what I'm seeing when I open my social media right now about this bill.

**SPEAKER_3** (1:06)
The Amber Mac Show on Canada Talks, SiriusXM 167

**Amber Mac** (1:12)
Welcome to The Amber Mac Show. Coming up in today's episode, how Canada plans to keep kids and maybe all of us safer from online, the rising costs of artificial intelligence and the tech leader who is now a trillionaire.

**Jeff MacArthur** (1:24)
And we'll also talk about the countries that want tech companies to share their wealth with workers, how to upskill the next generation, and Marc Saltzman joins us with some cool tech for dads and grads. But first, let's get to the headlines.

**Amber Mac** (1:39)
The Tapod D235 Video Doorbell Camera was recently awarded Tech Hive's Best of 2025 to 2026
That's no surprise since it does so much. First up, it uses smart AI to identify people, pets, vehicles, and packages, and notifies users as needed. Second up, you can save recorded videos to the micro SD card up from 512 gigabytes, it's sold separately, or use TapoCare Cloud Storage subscription required. Plus, the D235 exceeds traditional full HD resolution with finer details and incredible clarity.
In other words, you're getting a 180 degree ultra wide field of view, capturing every detail from head to toe in a single sweeping glance. You can order your TP-Link Video Doorbell Camera today over on Amazon. Thanks so much to TP-Link for supporting The Amber Mac Show.

**SPEAKER_3** (2:35)
Headlines.

**Jeff MacArthur** (2:37)
Well, after a brief flirtation with token maxing, we're shifting to the next era of AI where a lot less can be done for free, and the costs of doing more are finally becoming apparent as AI companies stop subsidizing usage.
From GitHub users complaining that just a few hours of AI use can take up most of their monthly subscription quota, to a company using Anthropics Cloud ending up with a half a billion dollar bill for a single month of usage, we're starting to see reactions like the recent one from Amazon senior vice president who was pleading with staff to quote, not use AI just for the sake of using AI, end quote. And let's be totally clear, this would be staff who were previously strongly encouraged by management like him to use as much AI as they could.

**Amber Mac** (3:27)
We should also add that OpenAI CEO Sam Altman has said that the growing costs of AI are becoming a problem and judging by the number of token overspending memes that I think we're both seeing on social media, he is far from the only one. Hopefully this serves Jeff as a wake up call for AI zealots that there is a limit to how useful AI can be, especially when its costs are compared with those of having a human worker doing the same task. In fact, a German court has ruled in a case that held Google accountable for the many wrong answers in Google's AI summaries that AI search is, and I'm quoting here, an additional function, one without which the use of the search engine will still be and is possible, and without which users are perfectly capable of finding results amidst the flood of data.

**Jeff MacArthur** (4:18)
Here, here. You know what I think of AI search from my previous The Other Mac segment on that? So this kind of pushback makes me very happy. The court went on to say that while traditional search results simply present lists of links to third-party information, Google's AI search makes independent, new, and substantive statements that can be the result of its own misinterpretation of online content, not some third-party claims that Google couldn't prevent even if they wanted to. In fact, The New York Times published an article about a month or so ago that pointed out that Google's AI overviews drew responses from untrustworthy sources like Facebook posts, resulting in incorrect statements in almost one in 10 of the AI results it serves up, as well as having inaccurate link sources more than half the time.