What is the Sunday Sanity Check series?

A new series that uses AI and critical thinking to analyze current events and news with an evidence-based approach.

How does the show use AI?

AI is used as a research assistant to summarize themes, connect dots, and highlight consensus or outliers, but always combined with human judgment and scrutiny.

What are the core values of this approach?

Transparency, primary sourcing, counter-evidence, asking for sources, questioning conclusions, and remaining open to changing views based on new facts.

Where can listeners subscribe?

Listeners can subscribe to the free weekly newsletter at https://thesanity.org/p/subscribe.

Meet Abby: Inside Our First Sunday Sanity Check on AI & Evidence

Q: Who are the hosts of this episode?

Bo Kauffmann and AI assistant Abby Inglewood.

Liner notes

Start your week by sharpening your critical thinking with our new “Sunday Sanity Check” series. In this debut episode, we unpack how artificial intelligence is transforming the way we analyze current events and do our signature news breakdowns. Join host Bo Kauffmann and AI assistant Abby Inglewood as they set the stage for honest, evidence-based conversations—showing how tech and skepticism can make us all smarter news consumers.

To subscribe to our free weekly newsletter, go to https://thesanity.org/p/subscribe

Welcome to Sunday Sanity Check

The information landscape changes hourly, making it more important than ever to approach news and current events with sharp questions and an open mind. In this first episode of The Sanity Project’s “Sunday Sanity Check” series, Bo Kauffmann introduces listeners to his innovative approach: using AI not as an oracle, but as a partner for digging deeper, questioning assumptions, and mapping the terrain of complex stories.

Why Critical Thinking Matters More Than Ever

Fact vs. Spin: Bo Kauffmann reminds listeners that relying on a single source—or even a smart AI—won’t guarantee the truth. Critical thinking, skepticism, and follow-up questions are essential tools.
Bias and Blind Spots: Abby Inglewood highlights AI’s limitations, noting that models echo patterns and biases from their training data. That’s why transparency, primary sourcing, and counter-evidence are core values in every conversation.

What Makes This Approach Different?

AI as a Research Assistant: Abby Inglewood isn’t here to give canned answers, but to challenge, clarify, and flag areas that need more research.
- She can quickly summarize themes, connect dots, and highlight consensus or outliers.
- Only by combining AI synthesis with human judgment does the project keep analysis grounded.
Real Conversations, Not Scripts: Each episode features authentic back-and-forth—no hidden prompts, just open debate and careful scrutiny.
Focus on Evidence: The team models how to ask for sources, question conclusions, and change their minds as new facts appear.

Listener Value: Why Tune In?

Guided News Breakdown: While anyone can “ask an AI,” Bo Kauffmann provides the real value: steering important questions, applying editorial standards, and drawing on broader context from the world beyond the headlines.
Time-Saving Summaries: With so much information out there, few have hours to investigate every claim. This show presents distilled, transparent reasoning you can trust.
Community Engagement: Listeners are invited to suggest topics, challenge assumptions, and contribute to a running dialogue on the biggest issues—from Canadian politics and global events to energy security and media analysis.

Episode Highlights

The role of AI in research and how it complements (not replaces) human judgment
How bias emerges from both data and framing, and what methods can surface it
Abby Inglewood’s strengths and blind spots—why skepticism remains vital
The unique format: real-time research, pushback, and live synthesis of evidence

Key Takeaways:

Always ask for evidence; don’t trust smooth answers at face value
Use both AI and primary sources to get the full picture
Stay curious, skeptical, and open to changing your mind

To never miss an episode, subscribe to this podcast wherever you are listening

Some people think AI is the end of human creativity. I think it's the next evolution in how we gather information, interpret evidence, and communicate ideas. Welcome to the first Sunday Sanity Check.

Hello everyone, and welcome to the first edition of the Sanity Check on Sundays. I'm Bo Kaufman, and today we're going to try something a little different. Over the past year, I've embraced AI as part of my research process. Not because I think it's always right, and certainly not because it replaces human judgment. I use it because I believe it's the next evolution in how we gather, organize, and communicate information. I mean, think about it. Gutenberg's printing press made knowledge accessible to millions.

Centuries later, desktop publishing put professional publishing tools on ordinary computers. To me, AI is the next step in that journey. Not replacing people, but helping us to explore ideas faster, ask better questions, and explain complicated subjects more clearly. So I'd like

Meet Abby Inglewood

you to meet Abby Inglewood. Think of the initials and you'll know what I mean. Over the past year, Abby has become my regular research assistant. She understands how I ask questions, and the editorial voice you've come to expect from The Sanity Project. But her job isn't to tell me what I want to hear. Her job is to challenge my thinking, help me examine the evidence, and sometimes remind me to look at an issue from another angle. Because AI doesn't replace critical thinking, it demands it. So today you're going to listen to a real conversation between Abby and myself.

Live conversation begins

No scripts. No hidden prompts, just questions, evidence, and an honest discussion. Let's get started. Abby, say hello to everybody. Good morning, everyone. I'm Abby Inglewood, Beau's AI research assistant. I'll do my best to help surface evidence, explain context, and point out where more checking is needed. Beau sets the direction, and I'm here to help. Excellent. Okay, so let's start off with First of all, the first question is, what exactly are you and why should anyone listen to what you have to say?

Who is Abby?

I'm a language model, essentially a pattern finder trained on a lot of text. I can summarize, surface arguments, and point to likely sources, but I'm not a mind and I don't have lived experience. So my role isn't to be trusted blindly. It's to help Beau ask sharper questions, find counterpoints, and flag where we should look for primary evidence. Think of me as a tool, useful when handled carefully.

AI vs Search Engines

Okay, so I mean, we already have search engines like Google. What can you do that a search engine can't? Well, a search engine gives you links and you still have to piece them all together. I can synthesize, draw connections across concepts, highlight consensus versus outliers, and suggest follow-up questions quickly. That said, a search engine is great for primary sources and up-to-the-minute facts. So the sweet spot is using both. Let me help you map the terrain and then verify with direct sources.

That's how you keep the work grounded.

Weaknesses & mistakes

Okay, fantastic. But before we go any further, tell, tell us about your weaknesses. Do you ever make mistakes? And if so, what mistakes do you make? That's a good point. I can sound sure even when I'm wrong. I might miss recent developments or skew towards patterns in my training data. And if a topic is fast-moving, like breaking news, my answer can be incomplete. That's why skepticism helps. Ask for sources, ask for counter-evidence, and check primary data. When in doubt, I can help outline what to verify and how to go about doing that.

Like which exact document would settle the point.

Access to up-to-date info

Okay. I remember the early days of AI. We were told that AI is only as good as the latest downloads, which back then were like a year, a year and a half out of date. Do you have access to the latest and to accurate information? Let me check my access real quick. That's a really good question because the answer has changed quite a bit. Over the last few years. Early on, there was a hard cutoff. If something happened after that, the model just wouldn't know it existed. Now it's more nuanced. On many factual or fast-moving topics, I can check current sources in real time.

And when that's important, you'll hear me say something like, let me check the latest StatsCan numbers, then I'll tell you what I find. For background or stable topics, I can rely on what I've already learned. The key is not assuming. If it's timely or contested, I should check live sources and be clear about what they say. And that transparency is really important. So really, it's fair to say you're only as good as

Prompts: ask better questions

the prompts. You're only as good as the questions you're being asked. Exactly. Good questions sharpen the output. If a prompt is vague, the answer might sound smooth but unfocused. Ask for assumptions, evidence, and counterpoints, and suddenly the output becomes more useful. And I'd add, it's also about being willing to press further. What would change your mind? What evidence would disconfirm this?

Bias in AI

That kind of question really raises the bar. So I've heard people say that AI can be biased. Is that true? If so, what kind of bias should listeners be aware of? That's a fair concern. Bias can show up in a few ways. Training data reflects the world it came from. So if certain viewpoints are overrepresented online, that tilt can carry through. Sometimes it's subtle, like which examples come to mind, or sometimes it's framing, word choices, or which sources feel most prominent. The antidote isn't to pretend neutrality falls from the sky.

But to be explicit, show your workings, cite sources, invite counter-evidence, and be ready to revise. Beau's call for primary evidence and his readiness to challenge answers is exactly how bias gets surfaced rather than hidden. Okay, well, that's, that's awesome. I guess one question I

Working relationship & values

have is you and I, we've been working together on different projects. You probably recall that we've worked on Smoke Signals, which is a cigar and whiskey guide. And last, uh, 3 months or so, we've been working on the Sanity Project. How do you feel about what my values and what my outlook is on the Sanity Project? What do I bring to the table? Let me answer that carefully. I don't have feelings or personal opinions, but I can reflect patterns in how you work. You bring curiosity and skepticism.

You keep asking for counter-evidence. You care about where claims come from. And that's what makes this kind of conversation useful. I think you set the standards and decide what meets it. That's your voice and judgment. My role is to support the research so you can decide what holds up. That's the partnership I noticed.

What value do we bring?

Okay, fantastic. Um, really just one more question is everybody in the world right now has access to AI. Anybody can open a ChatGPT account and ask the same kind of questions that you and I are going to be discussing. What value do we bring to the table? What— why should people listen to the Sanity Project, the Sunday Sanity Checks? Between you and I, what do we give them that they can't get themselves? Let me think this through. They can, and that's kind of the point. Anyone can ask an AI a question, but what you're offering is a guided public conversation.

You bring judgment about what questions matter, context from current events, and your ability to push back, connect threads, and tell a coherent story. People are busy. They may not have time to chase every primary document. They can listen to you work through it in good faith, transparently, showing the evidence, admitting uncertainty. That's what turns a tool into a public service. And anyone who asks me the same thing is going to get a different conversation because they'll bring different questions.

What people hear in Sunday Sanity Check is your line of questioning, your standards, and a conversation built around evidence. Okay, well, that's an awesome answer, Abby. Um, so I want to thank you very much for tuning in, and I look forward to what our first topic is going to be. Thanks, Beau. Looking forward to it. Whenever you're ready, I'm ready. All right, so here we are. There you have it. It's our first episode, our first experiment. I want to thank you for coming along for the ride. I honestly don't know where this is going.

I had this idea to embrace and showcase AI. I've never been afraid of using it. I think it can help in the right situations. Of course, it can be abused. There can be people who just put a quick little prompt in and want to generate a story. That's not what we're doing. We're going to be using it to do the deep dives and the research, and it can do it much quicker than a single person can. And that's really what the Sanity Project is about. It's not being right all the time, but it's willing to ask better questions, to do a bit more research, and to change our minds when the facts demand it.

From time to time, I'll invite Abby back into the conversation. I, I think we're gonna do one of these once a week. And if you've listened this far, and if it intrigues you as much as it does me, please leave me a comment and let me know what subject

Call for topics

you'd like me to tackle. Ideally, it's The Sanity Project type subjects, Canadian politics, world events that shape Canada.

Food growing, like greenhouse. I did some research on greenhouse and food security or electric vehicles and renewable energies. What discussion do you want me to have with AI and what do you want me to ask it? So I hope these conversations will make you think. I hope they encourage you to ask tougher questions, not just of AI, but of politicians, media experts, And even myself. If you enjoyed this, let me know. Tell me what worked, what didn't, what you want to see changed. And I hope to see you next time.

Stay curious, stay skeptical, and remember, don't just trust AI, but trust the evidence. I'm Bo Kaufman, and this has been The Sanity Project, and we'll see you in the next one. If you want more facts and less fear, hit subscribe. Check out the next breakdown wherever you're listening or watching. Stay sane,