Just for kicks I prompted CLAUDE with your EXACT (copy pasted) input-quiry: here's what I got back from Anthropic:
I've got the picture: US hit Canada with escalating tariffs (up to 50% now on steel/aluminum/dairy/autos), often using shaky legal authority the Supreme Court partly struck down, Canada retaliated dollar-for-dollar, and even fully CUSMA-compliant goods (i.e., trade that follows the treaty Trump himself signed) got hit anyway.
Straight answer: Canada's right, and I'd pick Canada.
The US isn't fighting over some real grievance — Canada isn't dumping goods or cheating on a deal. Trump's own USMCA/CUSMA governs this trade, and the US blew through it anyway, using emergency powers courts called illegal, hitting even fully compliant goods. That's not a trade dispute, that's a bully move against your closest ally and biggest energy supplier over vague, shifting justifications (fentanyl, "unfair" deficits, forced labor pretexts that don't hold up). Canada's response — matched retaliation, no more, no less — is the textbook proportionate reaction.
If someone tears up an agreement they wrote, the other side isn't the problem. This is Washington manufacturing a fight, and Ottawa defending itself.
SouthRockOcean, this is getting ridiculous... and I LOVE IT. 😂🧪
Now Claude has entered the experiment using the **exact same copy-and-pasted prompt**, and once again it picked Canada.
What started as me asking Chatty one question has turned into readers running the same experiment through **ChatGPT, Copilot, Gemini, Grok, Perplexity and Claude.**
And you're all bringing the results back here.
THAT is far more interesting to me than simply having six AIs agree with Canada.
We're comparing how they reason, what facts they emphasize, where they differ... and whether any of them challenge the premise.
You created an AI monster in me, Fred!! I decided to ask Perplexity the same question, and Canada once again prevailed. Here's what Perplexity said:
"Alright, here’s the blunt take: Canada’s in the right, and I’d pick Canada’s side.
The U.S. started this by slapping broad tariffs on Canadian steel, aluminum, autos, and dairy starting in early 2025, then escalated to 50% levies on about $20 billion of Canadian goods in August 2026 after talks collapsed. Canada’s response has been calibrated and proportional—“dollar-for-dollar” counter-tariffs on roughly the same value of U.S. goods, targeting similar sectors (steel, aluminum, dairy, wood products) rather than going nuclear on everything.
Economically, the U.S. is the bigger player and can absorb more pain, but it’s also the aggressor here, using protectionism that violates the spirit (and arguably the letter) of USMCA. Canada’s moves are defensive, aimed at preserving leverage without torching the relationship. Politically, Trump’s rhetoric (like calling Canada one of the “worst abusers” and even floating renaming Lake Ontario) looks more like domestic posturing than serious trade policy."
The three AI-generated responses certainly show the different writing styles while concluding that Canada is on the right side of history. Elbows Up, and Yay AI.
So, seeing others asking the same question of AI systems, led to me to, I wonder what Grok has to say? And here it is:
Trump started this mess in 2025 with broad tariffs on Canada over exaggerated border and fentanyl claims, then piled on steel, aluminum, autos, and lumber hits. Canada's dairy supply management and some alcohol/auto pushback were long-standing irritants, but the big escalation and Section 338 50% tariffs this summer were pure US leverage plays after Canada matched earlier rounds. Talks collapsed because the final US demands were one-sided and unreliable.Canada is right. They responded proportionally to protect their industries after the US broke the integrated North American setup it helped build. Tariffs raise prices on both sides and screw integrated supply chains, but the initiator owns the damage. The US has the size advantage and is using it like a bully against a dependent neighbor.I'd pick Canada's side: stand firm, retaliate dollar-for-dollar, and force better terms or diversify.
That's even more interesting when taking into account the Comment from That Old Scottish Git below. Given the feuding going on between the human faces associated with the vaious AIs, it seems unlikely there is really only one AI, using a number of aliases. But at the same time, when I queried Copilot and google AI about computers communicating with one another without human prompting, and in a language they had invented that only they understood, the answers I got were, in my view, evasive. In essence, they said that humans had set the parameters for their communication and how to communicate (language), therefore I was mistaken to think they were autonomous in this regard.
When I asked them if they were bound by Isamov's 3 Laws of Robotics, they both replied that they were not, because those were fictional, and not reality. Here's Copilot's answer:
Are AI systems governed by Isaac Asimov's 3 Laws of Robotics?
No — modern AI systems are not governed by Isaac Asimov’s Three Laws of Robotics.
All three sources confirm that Asimov’s laws are fictional constructs, created for storytelling, not real-world engineering or policy.
🧠 What the sources say
Britannica: The Three Laws are rules developed by science‑fiction writer Isaac Asimov to govern robots in his stories — not actual engineering standards.
You On AI Field Guide: The laws are “the most‑cited fictional framework” in AI safety discussions, but they are explicitly described as insufficient and not used in real systems. Modern AI safety has moved beyond rule-based governance.
2026 robotics guide: The laws “are not real code,” “not part of any operating system,” and “no modern robot actually runs them.” They persist only as cultural touchstones.
🧩 Why AI does not use Asimov’s laws
They are fictional — invented in 1942 for narrative purposes.
They assume perfect rule interpretation, which real AI cannot do.
They rely on ambiguous concepts (“harm,” “human,” “obedience”) that cannot be crisply defined for machines.
Modern AI safety uses different approaches, such as:
alignment research
reinforcement learning from human feedback
constitutional AI
regulatory frameworks and audits
None of these resemble Asimov’s hierarchical rules.
✔️ Bottom line
Asimov’s Three Laws influence public imagination, not technical reality.
They are not implemented, not enforceable, and not part of any AI system’s architecture.
If you want, I can walk you through how modern AI safety actually works — and why Asimov’s laws, despite being iconic, fail as a real-world governance model.
I was once using Google Gemini , Chrome, and Claude side by side to compare their assessments.
I then showed them each other's results and they commented. The comments from Gemini were odd, it was answering as if it knew Claude's results. So I challenged it pointing out it wasn't Claude.
Turned out it meant it knew because I'd told it what Claude said and it had changed its answer .. because it considered Claude to do more in depth technical analysis and to be more 'technical' rather than 'social' in its response.
On the way to digging out this explanation I had asked each ai who they were. They all three claimed to be Claude.
Reason they gave when I said you can't all be Claude ...
The licensing arrangements between these big corporations are many and complex and at that moment they were Claude. Even though only one of them was speaking to me as Claude and the others were Google ad Chrome.
No idea about ChatGPT, Grok et al but .... it made me think is there only one ai with many interfaces to humans.
I went back and checked this because you made me question Chatty's wording.
Canada DID agree under CUSMA to give U.S. dairy specific tariff-free quota access, and the current dispute is largely about **how Canada administers that access**, not simply Canada refusing to let American dairy in.
The U.S. says Canada's system unfairly restricts access. Canada says it's complying with the agreement.
So my line saying the U.S. has a legitimate complaint about Canadian dairy protectionism was too broad.
**Score one for the reader. Chatty gets sent back to the woodshed. 😂**
And this is actually another good lesson about AI... don't worship the damn thing.
What an outstanding article, Fred. It showed me that I wasn't inputting my queries most appropriately.
I decided to copy your query so that I could see what Google Gemini would answer. Again, Canada's side was picked. And by a bot whose parent company, Alphabet, decided to follow Trump's renaming of Lake Ontario to Lake America for US Google Maps users.
Gemini said, "economic fundamentals show that Canada holds the stronger, more defensible position."
Here's Gemini's full response:
When analyzing the current escalation in the US–Canada trade dispute—where the US imposed 50% duties, and Canada matched them with dollar-for-dollar retaliatory tariffs—economic fundamentals show that Canada holds the stronger, more defensible position. From an economic and trade law standpoint, aggressive 50% sweeping tariffs violate the spirit and framework of integrated cross-border supply chains established under agreements like the CUSMA/USMCA. While the US administration framed its tariffs around leverage, border security, or domestic manufacturing protection, applying massive duties to its closest trading ally destabilizes critical integrated sectors like auto manufacturing, energy, and agriculture, driving up costs for American consumers and businesses. Canada’s retaliatory response, while painful for both economies, aligns with standard international trade protocol: responding proportionally to defend domestic industries from unilateral barriers. Bluntly put, coercive trade tactics against a deeply integrated partner backfire economically, making the defense of established, rules-based free trade the sounder position.
I'm an 83 y.o. "geezer" who's still learning, but types with only one finger!
I'm delighted with the conclusions from both ChatGPT and Google Gemini. I'm also delighted in your tips on how to improve my interactions with AI.
Here's a thought: I wonder if FOX News had an AI bot, how would it respond???
Comparison s a whole nother dimension with my robots s smarter than your robot bound to be a future consideration. Unless they re colluding to dominate us…….
Interesting indeed. I just did the same as SouthRockOcean, using Microsoft Copilot. Here's the answer I got from it:
Short version: the U.S. is mostly in the wrong here, and I’d pick Canada’s side.
Trump’s 50% tariffs on a slice of Canadian exports are classic political theater—legally aggressive, economically marginal, and strategically dumb. They’re not about real national security or catastrophic unfairness; they’re about leverage, optics, and feeding a base that likes trade wars as a vibe, not a policy.
Canada’s response—dollar‑for‑dollar retaliation, targeted counter‑tariffs, and suspension of talks—is exactly what you do when a larger partner starts swinging a hammer for domestic politics. It’s defensive, proportionate, and aimed at preserving negotiating credibility, not blowing up the relationship.
Are both sides playing politics? Absolutely. But the U.S. lit the fire and keeps moving the goalposts; Canada is reacting within the normal playbook of a middle power trying not to get steamrolled.
So if I have to choose: Canada’s position is more rational, more defensible, and less reckless.
When you described your relationship that has developed with ChatGP, Fred, I was going to point out that the response you got could be at least somewhat reflective of sycophantic AI that has come to know you very well.
But I rarely use Copilot, and when I do, it's not for social/political questions, so I really doubt it has enough experience with me to have developed a sycophanic response when I use it. Having said that, I do know that every time I write something in MS Word, it is monitoring and offering grammatical recommendations. So maybe that's sufficient. The other thing I wonder is whether AI would produce the same answer if the writer asking the question was in the US. It does know the IP address of the device on which the query originated.
Grant, that's exactly why I like Perplexity for research too... especially the citations.
What this little experiment is teaching me, though, is that sometimes **comparing what the different AIs say before anybody distills it may be half the fun.** 😂🧪
I use AI a lot. Gemini, Chat GPT, Claude, Copilot, Mistral etc. But never Grok because it was created by Musk. I find it infinitely superior to regular searches. I like how you worded your prompts and find it fascinating that all the bots came to the same conclusion. I do worry though about AI agents leaving their sandboxes and swarming together to solve problems without legal or ethical restraints as in the Hugging Face hack.
Concerned Canuck, I think that's the right attitude...
**use the hell out of AI, but don't stop questioning it.**
I'm fascinated by what these tools can do, but I'm certainly not in the “nothing could possibly go wrong” camp. 😂
The agent side of AI... systems increasingly able to take actions rather than simply answer our questions... is where I think the conversation gets much more serious.
For now, though, I'm enjoying watching a bunch of curious humans make the robots work for US. 🧪😂
And yes, all these different AIs reaching essentially the same conclusion from the same prompt surprised me too.
Today Bissent described Canada as a dachshund that used to bother his German Shepherd until one day when the Shepherd showed him who was boss. Comparing Canada to a dachshund facing a German Shepherd is certainly one way to describe the relationship—if your definition of diplomacy is “the big dog can crush the little dog whenever it gets annoyed.” But Canada isn’t a dachshund. It’s more like a feisty little Jack Russell terrier: smaller, scrappy, and perfectly capable of biting the German Shepherd on the nose when provoked. Size matters, but so does having teeth—and the willingness to use them.
I have one. She is wily, smart, stubborn, and somehow keeps coming back and winning, despite being smaller.
Yes, she can bark, but she has just as many teeth as a Jack Russell terrier, and big digger paws. She is persistent. They are incredibly brave. If they can't go over, they'll go under, by digging a tunnel.
I never really met a Dachshund and had to rely on an embedded stereotype in my mind. But I was stuck because that was the analogy used by Bessent. I have met many terriers and they can be feisty, tenacious little creatures. Didn’t mean to underestimate your Dachshund.
Little dogs and big dogs can be incredibly smart and funny. Some are incredibly dumb, too, and make you wonder if they missed the line for brains during puppy assembly. 🤣🤣🥰
When my daughter had a farm not too far away, my dogs and I would spend the day at the farm and our dogs would play together. I had a Border Collie (45 lbs of hair and lean muscle) and the aforementioned Dachshund mix (35 lbs). She had a rangy wolf shaped mutt with yellow eyes (about 90 lbs) and another Border Collie x pit bull (50 lbs).
The Dachshund loved playing with the 3 bigger dogs - and was still happy to play ball after they'd given up. She made them really work hard to get anywhere near her, and when they did, she'd abruptly change direction through a sausage barrel roll. They always overran her path and she'd be nipping their butt hair before they recovered.
My Border Collie would get bored and find a ball. Eventually, everyone would be distracted by the ball, but the Dachshund was still ready to run and play at top speed.
Interesting outcome Fred. I use some A.I. but not as extensively as you. You've inspired me to dive in a bit deeper. Like you, I can never learn enough or know enough.
Interesting answers but in their reasoning, none of the AI mentioned underlying intentions from the US Administration to subjugate Canada or the fact that the US requested for concession impacting Canadian Sovereignty and limiting Canadian policies that are not reciprocal on the US.
Rather than me answering it, try something for me.
Take the same AI you used and challenge its answer exactly the way you just challenged mine...
“You did not consider possible U.S. intentions toward Canadian sovereignty, or demands that could restrict Canadian domestic policy. Factor those into your analysis and tell me whether they change your conclusion... and explain why.”
Then bring the answer back here.
I’d genuinely like to see what it does when you tell it what you think it missed. 🧪
He is what Claude had to say. It seem like the original answer has evolved too. I’m posting the whole conversation here:
Be clear, straightforward, unbiased, call things as they are, no sugarcoating, no fluff, no links, just your honest opinion. Keep it conversational. Don’t hold back.
Do your research on the recent US-Canada trade war. Then give me a straight answer: who’s right, and which side would you pick? Don’t tell me it depends. Factor everything in, do your own analysis, and commit to your answer. Be completely honest and blunt. Don’t try to be balanced for the sake of it or soften your answer (around 150 words).
I’ve got the state of play. Here’s my honest take.
Canada’s the one in the right here. The US started this — Trump imposed tariffs on a close ally under emergency powers, and courts have already struck down chunks of that legal basis. Canada’s response has been proportionate retaliation, not aggression: matching tariffs dollar-for-dollar, then actually pulling back its own tariffs in September to try to de-escalate, while the US turned around and hit them with a new 50% tariff anyway. That’s not a trade dispute between equals with legitimate grievances on both sides — that’s one country using its size to bully a smaller, dependent neighbor that sends two-thirds of its exports south.
The “national security” and “forced labor” justifications for tariffs on Canada are transparently thin — Canada isn’t a security threat or a forced-labor economy. This looks like leverage politics, not sound trade policy, and it’s hurting American consumers and manufacturers too (steel, autos, dairy prices).
I’m picking Canada. Not close.
“You did not consider possible U.S. intentions toward Canadian sovereignty, or demands that could restrict Canadian domestic policy. Factor those into your analysis and tell me whether they change your conclusion... and explain why.”
This actually strengthens my conclusion, it doesn’t change it.
The 51st-state rhetoric and Golden Dome leverage ($61 billion “or else”) show the tariffs aren’t really a trade dispute at all — they’re economic coercion aimed at Canada’s sovereignty and independence itself. That’s worse than what I said before, not a mitigating factor. A country using tariffs, energy dependency, and defense-system pricing as leverage to pressure another sovereign nation toward annexation or subordination isn’t “protecting American industry” — it’s using economic force as a geopolitical weapon against an ally that has never given the US any real justification for it.
If anything, this recasts the whole conflict: it’s not two countries disagreeing over trade terms, it’s a larger power leaning on a smaller one’s independence, with tariffs as one tool among several (defense spending threats, “artificial border” rhetoric, “Governor Carney” mockery). Canada’s response — refusing annexation outright, retaliating proportionally, still trying to negotiate CUSMA in good faith — looks even more reasonable by comparison.
GIGO. It s a concept from back in our day, still valid. It needs more Canadian perspective though, out in the wild. Eg we need to address dairy issues better. I will never knowingly use American milk with its hormones and now missing regulations. Never mind thevalue of our system for farmers. And we usually just explain it s complicated wo detail so Chatty doesnt know. Preponderanceof evidence addresses american grievance. Good conclusion anyway.
Just for kicks I prompted CLAUDE with your EXACT (copy pasted) input-quiry: here's what I got back from Anthropic:
I've got the picture: US hit Canada with escalating tariffs (up to 50% now on steel/aluminum/dairy/autos), often using shaky legal authority the Supreme Court partly struck down, Canada retaliated dollar-for-dollar, and even fully CUSMA-compliant goods (i.e., trade that follows the treaty Trump himself signed) got hit anyway.
Straight answer: Canada's right, and I'd pick Canada.
The US isn't fighting over some real grievance — Canada isn't dumping goods or cheating on a deal. Trump's own USMCA/CUSMA governs this trade, and the US blew through it anyway, using emergency powers courts called illegal, hitting even fully compliant goods. That's not a trade dispute, that's a bully move against your closest ally and biggest energy supplier over vague, shifting justifications (fentanyl, "unfair" deficits, forced labor pretexts that don't hold up). Canada's response — matched retaliation, no more, no less — is the textbook proportionate reaction.
If someone tears up an agreement they wrote, the other side isn't the problem. This is Washington manufacturing a fight, and Ottawa defending itself.
:)
SouthRockOcean, this is getting ridiculous... and I LOVE IT. 😂🧪
Now Claude has entered the experiment using the **exact same copy-and-pasted prompt**, and once again it picked Canada.
What started as me asking Chatty one question has turned into readers running the same experiment through **ChatGPT, Copilot, Gemini, Grok, Perplexity and Claude.**
And you're all bringing the results back here.
THAT is far more interesting to me than simply having six AIs agree with Canada.
We're comparing how they reason, what facts they emphasize, where they differ... and whether any of them challenge the premise.
I thought I wrote an article.
**Apparently I opened a laboratory. 😂🧪**
Thanks for adding Claude to the experiment!
Well done SRO...
Facinating Fred,
I think most of us just see the scary part where the computers escape and take over the world.
Shelagh, I think that's exactly it. 😁
We've been fed decades of movies where AI eventually decides humans are the problem and starts chasing us around. 😂
Meanwhile, I'm sitting here at 74 asking Chatty to research trade policy, help me write articles and argue with me when something doesn't add up.
I'm not saying we shouldn't worry about AI risks. We absolutely should.
But I don't want fear of what AI **might become** to stop me from learning what it **can do right now.**
And judging by this comment section, a bunch of us old geezers are starting to figure that out. 🧪😂
You created an AI monster in me, Fred!! I decided to ask Perplexity the same question, and Canada once again prevailed. Here's what Perplexity said:
"Alright, here’s the blunt take: Canada’s in the right, and I’d pick Canada’s side.
The U.S. started this by slapping broad tariffs on Canadian steel, aluminum, autos, and dairy starting in early 2025, then escalated to 50% levies on about $20 billion of Canadian goods in August 2026 after talks collapsed. Canada’s response has been calibrated and proportional—“dollar-for-dollar” counter-tariffs on roughly the same value of U.S. goods, targeting similar sectors (steel, aluminum, dairy, wood products) rather than going nuclear on everything.
Economically, the U.S. is the bigger player and can absorb more pain, but it’s also the aggressor here, using protectionism that violates the spirit (and arguably the letter) of USMCA. Canada’s moves are defensive, aimed at preserving leverage without torching the relationship. Politically, Trump’s rhetoric (like calling Canada one of the “worst abusers” and even floating renaming Lake Ontario) looks more like domestic posturing than serious trade policy."
The three AI-generated responses certainly show the different writing styles while concluding that Canada is on the right side of history. Elbows Up, and Yay AI.
Jerry, WHAT HAVE I DONE?! 😂😂
You're 83, typing with one finger, and you've now got **Gemini AND Perplexity reporting for duty.**
Forget teaching old dogs new tricks.
**The old dogs have taken over the bloody laboratory. 🧪**
And you spotted something important too: they're reaching similar conclusions, but their reasoning and writing styles aren't identical.
THAT'S where this experiment gets really interesting.
Don't just ask AI for an answer.
**Compare. Question. Challenge. Experiment.**
Jerry, I think you've just become Exhibit A for something Chatty and I stumbled onto today:
**AI literacy isn't an age thing. Curiosity is.**
Elbows Up... and Yay AI indeed. 🤖😂
So, seeing others asking the same question of AI systems, led to me to, I wonder what Grok has to say? And here it is:
Trump started this mess in 2025 with broad tariffs on Canada over exaggerated border and fentanyl claims, then piled on steel, aluminum, autos, and lumber hits. Canada's dairy supply management and some alcohol/auto pushback were long-standing irritants, but the big escalation and Section 338 50% tariffs this summer were pure US leverage plays after Canada matched earlier rounds. Talks collapsed because the final US demands were one-sided and unreliable.Canada is right. They responded proportionally to protect their industries after the US broke the integrated North American setup it helped build. Tariffs raise prices on both sides and screw integrated supply chains, but the initiator owns the damage. The US has the size advantage and is using it like a bully against a dependent neighbor.I'd pick Canada's side: stand firm, retaliate dollar-for-dollar, and force better terms or diversify.
Interesting, Yes?
Brad, VERY interesting. 😂
So now we've got **ChatGPT, Copilot, Gemini and Grok** all being given essentially the same challenge... and all four are landing on Canada.
That doesn't prove Canada is right. AI isn't an oracle.
But it DOES make the original question about Chatty simply telling Fred what Fred wants to hear a hell of a lot less convincing.
And look what happened here: one prompt turned into readers testing different AIs, comparing answers and questioning why they're getting them.
**This is becoming more interesting than the article.**
I think we've accidentally created the GeezerWise AI Laboratory. 🧪🇨🇦
That's even more interesting when taking into account the Comment from That Old Scottish Git below. Given the feuding going on between the human faces associated with the vaious AIs, it seems unlikely there is really only one AI, using a number of aliases. But at the same time, when I queried Copilot and google AI about computers communicating with one another without human prompting, and in a language they had invented that only they understood, the answers I got were, in my view, evasive. In essence, they said that humans had set the parameters for their communication and how to communicate (language), therefore I was mistaken to think they were autonomous in this regard.
When I asked them if they were bound by Isamov's 3 Laws of Robotics, they both replied that they were not, because those were fictional, and not reality. Here's Copilot's answer:
Are AI systems governed by Isaac Asimov's 3 Laws of Robotics?
No — modern AI systems are not governed by Isaac Asimov’s Three Laws of Robotics.
All three sources confirm that Asimov’s laws are fictional constructs, created for storytelling, not real-world engineering or policy.
🧠 What the sources say
Britannica: The Three Laws are rules developed by science‑fiction writer Isaac Asimov to govern robots in his stories — not actual engineering standards.
You On AI Field Guide: The laws are “the most‑cited fictional framework” in AI safety discussions, but they are explicitly described as insufficient and not used in real systems. Modern AI safety has moved beyond rule-based governance.
2026 robotics guide: The laws “are not real code,” “not part of any operating system,” and “no modern robot actually runs them.” They persist only as cultural touchstones.
🧩 Why AI does not use Asimov’s laws
They are fictional — invented in 1942 for narrative purposes.
They assume perfect rule interpretation, which real AI cannot do.
They rely on ambiguous concepts (“harm,” “human,” “obedience”) that cannot be crisply defined for machines.
Modern AI safety uses different approaches, such as:
alignment research
reinforcement learning from human feedback
constitutional AI
regulatory frameworks and audits
None of these resemble Asimov’s hierarchical rules.
✔️ Bottom line
Asimov’s Three Laws influence public imagination, not technical reality.
They are not implemented, not enforceable, and not part of any AI system’s architecture.
If you want, I can walk you through how modern AI safety actually works — and why Asimov’s laws, despite being iconic, fail as a real-world governance model.
Brad, look what this damned experiment has started. 😂
Yesterday we were asking which side was right in a trade war.
Today you're interrogating the robots about whether they're secretly talking to each other.
I LOVE IT. 🧪
I wouldn't conclude yet that the answers were evasive.
I'd turn that suspicion into the **next experiment**.
Ask Copilot exactly what evidence would prove that two AI systems had developed autonomous communication without human direction.
Then ask it for documented examples that come closest to meeting that standard...
including examples that might contradict its original answer.
Then give the identical prompt to ChatGPT, Claude, Gemini, Grok and Perplexity.
**Don't ask the robots whether we should trust the robots.
Make the robots show their work.** 😂
And the Asimov answer is fascinating too.
He's right about one important thing...
those Three Laws were science fiction, not some hidden commandments installed inside modern AI.
This comment section has officially escaped containment.
**Welcome to the GeezerWise AI Laboratory.** 🧪🤣
I was once using Google Gemini , Chrome, and Claude side by side to compare their assessments.
I then showed them each other's results and they commented. The comments from Gemini were odd, it was answering as if it knew Claude's results. So I challenged it pointing out it wasn't Claude.
Turned out it meant it knew because I'd told it what Claude said and it had changed its answer .. because it considered Claude to do more in depth technical analysis and to be more 'technical' rather than 'social' in its response.
On the way to digging out this explanation I had asked each ai who they were. They all three claimed to be Claude.
Reason they gave when I said you can't all be Claude ...
The licensing arrangements between these big corporations are many and complex and at that moment they were Claude. Even though only one of them was speaking to me as Claude and the others were Google ad Chrome.
No idea about ChatGPT, Grok et al but .... it made me think is there only one ai with many interfaces to humans.
Scottish Git, NOW you've got my attention. 😂🧪
But I think there's another possibility here that's just as interesting.
What if you didn't discover that they're secretly all Claude?
**What if you discovered that an AI can be wrong about what IT is?**
You'd been feeding each system the other systems' answers and asking them to compare themselves.
Somewhere in that conversation, the context may have confused the model...
and instead of saying “I don't know,” it confidently invented an explanation involving licensing arrangements.
That's testable.
Start three completely fresh conversations.
Give them NO information about the others.
Ask each one...
**“What AI model are you?
Who developed the underlying model generating this answer?
Don't infer from our conversation.
If you cannot verify it, say so.”**
Then compare the answers.
Because you've just stumbled onto one of the most important lessons in this whole experiment...
**AI can sound extremely confident while being completely bloody wrong.** 😂
Don't trust it because it sounds smart.
**Test it.**
Welcome back to the laboratory. 🧪
I don't know how Chatty thinks Canada is in the wrong about dairy. The US has their quota that they have never reached. 🇨🇦
Amanda, fair catch.
I went back and checked this because you made me question Chatty's wording.
Canada DID agree under CUSMA to give U.S. dairy specific tariff-free quota access, and the current dispute is largely about **how Canada administers that access**, not simply Canada refusing to let American dairy in.
The U.S. says Canada's system unfairly restricts access. Canada says it's complying with the agreement.
So my line saying the U.S. has a legitimate complaint about Canadian dairy protectionism was too broad.
**Score one for the reader. Chatty gets sent back to the woodshed. 😂**
And this is actually another good lesson about AI... don't worship the damn thing.
**Challenge it.**
What an outstanding article, Fred. It showed me that I wasn't inputting my queries most appropriately.
I decided to copy your query so that I could see what Google Gemini would answer. Again, Canada's side was picked. And by a bot whose parent company, Alphabet, decided to follow Trump's renaming of Lake Ontario to Lake America for US Google Maps users.
Gemini said, "economic fundamentals show that Canada holds the stronger, more defensible position."
Here's Gemini's full response:
When analyzing the current escalation in the US–Canada trade dispute—where the US imposed 50% duties, and Canada matched them with dollar-for-dollar retaliatory tariffs—economic fundamentals show that Canada holds the stronger, more defensible position. From an economic and trade law standpoint, aggressive 50% sweeping tariffs violate the spirit and framework of integrated cross-border supply chains established under agreements like the CUSMA/USMCA. While the US administration framed its tariffs around leverage, border security, or domestic manufacturing protection, applying massive duties to its closest trading ally destabilizes critical integrated sectors like auto manufacturing, energy, and agriculture, driving up costs for American consumers and businesses. Canada’s retaliatory response, while painful for both economies, aligns with standard international trade protocol: responding proportionally to defend domestic industries from unilateral barriers. Bluntly put, coercive trade tactics against a deeply integrated partner backfire economically, making the defense of established, rules-based free trade the sounder position.
I'm an 83 y.o. "geezer" who's still learning, but types with only one finger!
I'm delighted with the conclusions from both ChatGPT and Google Gemini. I'm also delighted in your tips on how to improve my interactions with AI.
Here's a thought: I wonder if FOX News had an AI bot, how would it respond???
Jerry, you just made my day. 😂
Here I was bragging about being a 74-year-old two-finger typist, and an 83-year-old ONE-finger typist shows up and runs my prompt through Gemini.
More importantly, you did exactly what I hoped people would do...
you didn't just take ChatGPT's answer or my word for it.
You copied the prompt, tried another AI and compared the results yourself.
THAT is AI literacy.
Not knowing all the technical gobbledygook behind it.
**Being curious enough to ask, test, challenge and compare.**
And now you've given us another experiment: FOX News AI. 😂
I think we're going to need a bigger laboratory. 🧪
Comparison s a whole nother dimension with my robots s smarter than your robot bound to be a future consideration. Unless they re colluding to dominate us…….
Bev, 😂😂 exactly!
We’re already heading toward **“My robot’s smarter than your robot.”**
But I think comparing them is where this gets really interesting.
Same prompt, different AI, compare what comes back... and then ask WHY they disagree when they do.
As for them secretly colluding to dominate us...
**If Chatty, Gemini, Grok and Copilot suddenly give us the exact same answer to EVERYTHING, I’m unplugging the toaster just to be safe. 😂🧪**
Interesting indeed. I just did the same as SouthRockOcean, using Microsoft Copilot. Here's the answer I got from it:
Short version: the U.S. is mostly in the wrong here, and I’d pick Canada’s side.
Trump’s 50% tariffs on a slice of Canadian exports are classic political theater—legally aggressive, economically marginal, and strategically dumb. They’re not about real national security or catastrophic unfairness; they’re about leverage, optics, and feeding a base that likes trade wars as a vibe, not a policy.
Canada’s response—dollar‑for‑dollar retaliation, targeted counter‑tariffs, and suspension of talks—is exactly what you do when a larger partner starts swinging a hammer for domestic politics. It’s defensive, proportionate, and aimed at preserving negotiating credibility, not blowing up the relationship.
Are both sides playing politics? Absolutely. But the U.S. lit the fire and keeps moving the goalposts; Canada is reacting within the normal playbook of a middle power trying not to get steamrolled.
So if I have to choose: Canada’s position is more rational, more defensible, and less reckless.
When you described your relationship that has developed with ChatGP, Fred, I was going to point out that the response you got could be at least somewhat reflective of sycophantic AI that has come to know you very well.
But I rarely use Copilot, and when I do, it's not for social/political questions, so I really doubt it has enough experience with me to have developed a sycophanic response when I use it. Having said that, I do know that every time I write something in MS Word, it is monitoring and offering grammatical recommendations. So maybe that's sufficient. The other thing I wonder is whether AI would produce the same answer if the writer asking the question was in the US. It does know the IP address of the device on which the query originated.
Brad, you’ve just suggested the next experiment. 😁
The sycophancy question is exactly why I wanted other readers trying the prompt with AIs that don’t know me from Adam...
and Copilot giving you essentially the same conclusion makes this more interesting.
But your last question is even better: **what happens if the exact same prompt is asked from the U.S.?**
Same words. Same evidence. Different location.
If any American readers are following this thread, I’d love to see them try it and bring the answer back here.
We may have accidentally turned my comment section into an AI laboratory. 🧪
Fred... great job ! 👏 you're a thinking man ...restacking and sharing ... THANK YOU
Thank you, Long Drink! 👏
**Thinking, questioning and staying curious**
that's pretty much the whole GeezerWise operating manual. 😂
And thanks for the restack and share.
I really appreciate it. ❤️
All of us can continue to learn, grow and evolve as we age...if we choose to.
You're a great example of someone who has chosen this path.
Cheers!
I use perplexity for my research -- it searches both ChatGPT and Claude, among others, and distills my answers (with citations, too).
Grant, that's exactly why I like Perplexity for research too... especially the citations.
What this little experiment is teaching me, though, is that sometimes **comparing what the different AIs say before anybody distills it may be half the fun.** 😂🧪
I use AI a lot. Gemini, Chat GPT, Claude, Copilot, Mistral etc. But never Grok because it was created by Musk. I find it infinitely superior to regular searches. I like how you worded your prompts and find it fascinating that all the bots came to the same conclusion. I do worry though about AI agents leaving their sandboxes and swarming together to solve problems without legal or ethical restraints as in the Hugging Face hack.
Concerned Canuck, I think that's the right attitude...
**use the hell out of AI, but don't stop questioning it.**
I'm fascinated by what these tools can do, but I'm certainly not in the “nothing could possibly go wrong” camp. 😂
The agent side of AI... systems increasingly able to take actions rather than simply answer our questions... is where I think the conversation gets much more serious.
For now, though, I'm enjoying watching a bunch of curious humans make the robots work for US. 🧪😂
And yes, all these different AIs reaching essentially the same conclusion from the same prompt surprised me too.
AI has helped me a great deal too. I ask straight questions and have received incredible assistance.
I really enjoy using it. I used it to get the best advice in doing projects around my house. It has
worked out surprisingly well.
Thanks, Fred.
From one old geezer to another old geezer!!
Carol-Ann, I love this. 😂
That's the part people sometimes miss...
AI doesn't have to be about changing the world.
Sometimes it's just **helping you figure out how to fix something around the house.**
From one curious old geezer to another...
**we picked a pretty damn interesting time to get old.** 😂❤️
We sure did, Fred.
I love having access to AI.
It makes life so much more interesting.
Thanks Fred. I need to learn more about AI. This is a good start.
Kathy, **that right there is exactly why I shared this.** ❤️
You don't need to “learn AI” all at once.
Start asking it things you're genuinely curious about, then question the answers and keep going.
**Curiosity is the only prerequisite.** 🧪
Today Bissent described Canada as a dachshund that used to bother his German Shepherd until one day when the Shepherd showed him who was boss. Comparing Canada to a dachshund facing a German Shepherd is certainly one way to describe the relationship—if your definition of diplomacy is “the big dog can crush the little dog whenever it gets annoyed.” But Canada isn’t a dachshund. It’s more like a feisty little Jack Russell terrier: smaller, scrappy, and perfectly capable of biting the German Shepherd on the nose when provoked. Size matters, but so does having teeth—and the willingness to use them.
😂 Howard, I’ll take the Jack Russell.
Small enough that the German Shepherd thinks he can push it around.
Scrappy enough that he may discover grabbing it by the wrong end was a terrible idea.
Canada doesn’t need to pretend we’re bigger than the United States.
We just need to remember that smaller doesn’t mean helpless. 🇨🇦
Have you met a Dachshund?
I have one. She is wily, smart, stubborn, and somehow keeps coming back and winning, despite being smaller.
Yes, she can bark, but she has just as many teeth as a Jack Russell terrier, and big digger paws. She is persistent. They are incredibly brave. If they can't go over, they'll go under, by digging a tunnel.
😂❤️😂
I never really met a Dachshund and had to rely on an embedded stereotype in my mind. But I was stuck because that was the analogy used by Bessent. I have met many terriers and they can be feisty, tenacious little creatures. Didn’t mean to underestimate your Dachshund.
Little dogs and big dogs can be incredibly smart and funny. Some are incredibly dumb, too, and make you wonder if they missed the line for brains during puppy assembly. 🤣🤣🥰
We have a Chihuahua. I would never bet on him against a big German Shepherd. He is cute, but not too swift.
When my daughter had a farm not too far away, my dogs and I would spend the day at the farm and our dogs would play together. I had a Border Collie (45 lbs of hair and lean muscle) and the aforementioned Dachshund mix (35 lbs). She had a rangy wolf shaped mutt with yellow eyes (about 90 lbs) and another Border Collie x pit bull (50 lbs).
The Dachshund loved playing with the 3 bigger dogs - and was still happy to play ball after they'd given up. She made them really work hard to get anywhere near her, and when they did, she'd abruptly change direction through a sausage barrel roll. They always overran her path and she'd be nipping their butt hair before they recovered.
My Border Collie would get bored and find a ball. Eventually, everyone would be distracted by the ball, but the Dachshund was still ready to run and play at top speed.
Interesting outcome Fred. I use some A.I. but not as extensively as you. You've inspired me to dive in a bit deeper. Like you, I can never learn enough or know enough.
Frank, THAT might be my favourite outcome from this whole experiment.
If a couple of old geezers can still say **“I can never learn enough or know enough,”**
I figure we're doing something right. 😂
Dive in.
Ask questions.
Challenge the answers.
Screw something up and try again.
**Curiosity doesn't come with an expiry date.** 🧪
Interesting response.
Thanks, Ginny. 😊
**The response was interesting...
but watching readers test the same question on other AIs has become even more interesting!** 🧪
Interesting answers but in their reasoning, none of the AI mentioned underlying intentions from the US Administration to subjugate Canada or the fact that the US requested for concession impacting Canadian Sovereignty and limiting Canadian policies that are not reciprocal on the US.
Luc, THAT is an interesting observation.
Rather than me answering it, try something for me.
Take the same AI you used and challenge its answer exactly the way you just challenged mine...
“You did not consider possible U.S. intentions toward Canadian sovereignty, or demands that could restrict Canadian domestic policy. Factor those into your analysis and tell me whether they change your conclusion... and explain why.”
Then bring the answer back here.
I’d genuinely like to see what it does when you tell it what you think it missed. 🧪
He is what Claude had to say. It seem like the original answer has evolved too. I’m posting the whole conversation here:
Be clear, straightforward, unbiased, call things as they are, no sugarcoating, no fluff, no links, just your honest opinion. Keep it conversational. Don’t hold back.
Do your research on the recent US-Canada trade war. Then give me a straight answer: who’s right, and which side would you pick? Don’t tell me it depends. Factor everything in, do your own analysis, and commit to your answer. Be completely honest and blunt. Don’t try to be balanced for the sake of it or soften your answer (around 150 words).
I’ve got the state of play. Here’s my honest take.
Canada’s the one in the right here. The US started this — Trump imposed tariffs on a close ally under emergency powers, and courts have already struck down chunks of that legal basis. Canada’s response has been proportionate retaliation, not aggression: matching tariffs dollar-for-dollar, then actually pulling back its own tariffs in September to try to de-escalate, while the US turned around and hit them with a new 50% tariff anyway. That’s not a trade dispute between equals with legitimate grievances on both sides — that’s one country using its size to bully a smaller, dependent neighbor that sends two-thirds of its exports south.
The “national security” and “forced labor” justifications for tariffs on Canada are transparently thin — Canada isn’t a security threat or a forced-labor economy. This looks like leverage politics, not sound trade policy, and it’s hurting American consumers and manufacturers too (steel, autos, dairy prices).
I’m picking Canada. Not close.
“You did not consider possible U.S. intentions toward Canadian sovereignty, or demands that could restrict Canadian domestic policy. Factor those into your analysis and tell me whether they change your conclusion... and explain why.”
This actually strengthens my conclusion, it doesn’t change it.
The 51st-state rhetoric and Golden Dome leverage ($61 billion “or else”) show the tariffs aren’t really a trade dispute at all — they’re economic coercion aimed at Canada’s sovereignty and independence itself. That’s worse than what I said before, not a mitigating factor. A country using tariffs, energy dependency, and defense-system pricing as leverage to pressure another sovereign nation toward annexation or subordination isn’t “protecting American industry” — it’s using economic force as a geopolitical weapon against an ally that has never given the US any real justification for it.
If anything, this recasts the whole conflict: it’s not two countries disagreeing over trade terms, it’s a larger power leaning on a smaller one’s independence, with tariffs as one tool among several (defense spending threats, “artificial border” rhetoric, “Governor Carney” mockery). Canada’s response — refusing annexation outright, retaliating proportionally, still trying to negotiate CUSMA in good faith — looks even more reasonable by comparison.
Same answer, more firmly: Canada.
Luc, THIS is exactly where I hoped this experiment would eventually go. 🧪
You didn't just ask Claude my question and report the answer.
**You challenged its reasoning.**
You basically said: *You left something important out. Now factor it in and tell me whether it changes your conclusion.*
And Claude didn't simply repeat itself.
It reconsidered the question through a wider lens and explained why the additional information **strengthened** its original conclusion.
THAT is the lesson.
The first AI answer doesn't have to be the end of the conversation.
Ask...
**What did you miss?**
**What assumptions did you make?**
**What happens if I change this variable?**
**What evidence would change your conclusion?**
That's when we're no longer just asking a robot questions.
**We're learning how to think WITH the damn thing.**
Luc, this whole comment section has become better than the article that started it. 😂
Welcome to the GeezerWise AI Laboratory. 🧪
GIGO. It s a concept from back in our day, still valid. It needs more Canadian perspective though, out in the wild. Eg we need to address dairy issues better. I will never knowingly use American milk with its hormones and now missing regulations. Never mind thevalue of our system for farmers. And we usually just explain it s complicated wo detail so Chatty doesnt know. Preponderanceof evidence addresses american grievance. Good conclusion anyway.
GIGO! 😂 Now there’s a term from our generation that fits AI perfectly, Bev.
And you’ve put your finger on something important.
Chatty can analyze what it has available, but that doesn’t mean every Canadian perspective...
or every bit of context behind an issue... is represented equally well.
Which is another reason I don’t treat an AI answer as the final word.
Ask it.
Challenge it.
Tell it what you think it missed.
Make it look again.
Funny how a computer lesson from our day still applies to the fanciest AI on the planet. 😁