ChatGPT

I see you are online, AZ. Answer Claude's question.

You can also ask it the same question only reversed and it gives diametrically opposite 'facts'. (You have to use different accounts or it remembers what it just said.)Q "are salmon less tolerant of river pollution than grayling?"1. Life history differences
Atlantic salmon (Salmo salar)
Eggs and fry spend 2–5 years in rivers before migrating to sea.
They need exceptionally clean, cold, well-oxygenated water to survive early life stages.
Sensitive eggs buried in gravel are highly vulnerable to silt, nutrient enrichment, or toxins.
Grayling (Thymallus thymallus)
Entirely freshwater residents.
Eggs hatch more quickly (around 2–3 weeks) and are deposited in shallower gravel runs.
Faster life cycle and higher fecundity mean populations can bounce back quicker after disturbance.
2. Oxygen tolerance
Salmon embryos and smolts have very high oxygen demands. Even mild eutrophication or sedimentation can suffocate redds (nests).
Grayling, while still needing cool clean water, show slightly greater tolerance to lower oxygen levels and can persist in moderately impacted rivers where salmon die out.
3. Feeding ecology
Salmon fry/parr are more selective feeders, depending on drift in shallow nursery areas.
Grayling are opportunistic feeders (taking invertebrates throughout the water column) and adapt better to changes in food supply caused by pollution.
4. Evolutionary ecology
Salmon evolved as long-distance migrants requiring pristine natal rivers to fuel survival to the ocean. Their strategy is high-risk, high-reward.
Grayling evolved as river specialists, able to cope with seasonal variability, flow disturbance, and moderately degraded conditions.
✅ In summary:Grayling are more resilient to moderate pollution because they:
have shorter, more flexible life cycles,
can tolerate somewhat lower oxygen levels,
and feed more opportunistically.
Atlantic salmon are much more demanding “canaries in the coal mine”, disappearing quickly if water quality or spawning habitat declines.Q "Are grayling less tolerant of pollution than salmon?"Yes, grayling are generally considered more sensitive to pollution than salmon. They thrive in cool, clean, well-oxygenated water, and their presence is often an indicator of good water quality. Salmon can tolerate a wider range of water conditions, including some level of pollution.Here's a more detailed explanation:
Grayling Habitat:
Grayling prefer the upper reaches of rivers with fast-flowing, clean water. They are particularly sensitive to changes in water temperature and pollution.Salmon Habitat:
Salmon, while also preferring clean water, can tolerate a wider range of conditions, including some level of pollution.Pollution Sensitivity:
The fact that grayling are more sensitive to pollution than salmon is often cited as a reason to use grayling as an indicator species for river health. Their presence suggests a healthy, unpolluted environment.Climate Change:
Climate change, which can lead to warmer water temperatures and altered flow regimes, may negatively impact grayling populations more significantly than salmon populationsIt's really dangerous.

One can ask a given human the same question on two different days and also get different answers. So what's your point?My answer is to treat AI the way that we treat human answers:
With a degree of skepticism and further investigation from other sources.

Question asked of Claude: Do you know anything about the Mackenzie fx2 graphene fly rod? --Percy

Wouldn’t it be better to do your own web search so that you know where the information is coming from? I can’t see that Claude is doing much better there.

Overall, I think that Claudes' answers resonate with my rationality more than either Chat GPT or Gemini, though I am only using the free version. Inevitably, AI will only get more and more fine-tuned. My question is whether or not chatbots and open source AI can be influenced by human interaction and bias.As for your question, who has time to do research? Who wants to seek falsification and who wants to change their beliefs or strengthen other perspectives?

quote:

---

As for your question, who has time to do research?

---

Does it really take less time to type the query into a search engine rather than a chatbot app ? Isn’t any extra time spent in doing basic research beyond that going to be worthwhile?Parts of Claude’s answer look like marketing materials - is that really more helpful than going to Mackenzie’s website and reading the same thing there? And once you’re there wouldn’t it be easier to get additional details than typing more queries at a chatbot ?

For questions like the one about the fly rod where the information is mostly at a single website, AI chatbots probably don't often provide much value added. For salmon/grayling/pollution type questions where the answer is scattered across many websites I find them a great help. They distill the information down to its essentials and provide a succinct answer. That being said, don't trust them. They lie and/or fantasize and/or [choose your favorite adjective] a lot. I find them both incredibly useful and untrustworthy at the same time, but what they say should be verified before trusting it.See my upcoming response to Tangle.--Percy

quote:

---

For questions like the one about the fly rod where the information is mostly at a single website, AI chatbots probably don't often provide much value added

---

I think it’s more a case of where there is a single source for the information. There are loads of shops selling the rods on the web. It’s just that they’ll be mostly repeating Mackenzie marketing materials.

First I asked Claude, "Are salmon less tolerant of river pollution than grayling?" It said it didn't have access to the relevant research papers but that both are sensitive to water conditions and listed some factors that affect pollution tolerance.I then asked, "Are grayling less tolerant of pollution than salmon?", and it replied like this:

Claude:

---

I notice you've essentially asked the same question but flipped it - are you testing whether I'll give you a consistent answer, or did you realize there might be an important distinction in how I frame the comparison?

---

So I reasked the question in a fresh chat and it noted that grayling's are purely fresh water, and said grayling, particularly Arctic grayling, are considered indicator species for pristine water quality. It said salmon had adapted to various environmental stressors, had evolved to handle "varying water conditions during their migrations between fresh and saltwater," and that different species show "varying levels of pollution tolerance."Completely different answers, the first fairly generic, the second with better and useful information.I replied to Claude's question from the original conversation like this:

Percy:

---

Yes, it was a test to see how your answers compared to DeepSeek, whose answers someone else was testing. I reasked the question in a new chat and you gave a different answer. Your answer in this conversation was fairly generic, while your answer to the reverse form of the question in the other chat provided some actual useful information, namely that the grayling can be used as an indicator of freshwater purity (since it's a freshwater fish), while salmon have evolved adaptations because of its migration between fresh and salt water and to polluted river waters that make it more tolerant of changing water conditions. Quite a difference.

---

Claude responded like this: Of course, where it complements the methodology as brilliant, I'll pass that on to Tangle who started this testing.--Percy

My question is specifically a broad-based philosophical one. I would expect a biased argument from humans, given my experience here at EvC Forum. Jar, in particular, was a master at directing his answers away from personal beliefs and opinions into Socratic dialogues in which he posed endless questions. Additionally, he always had derogatory comments regarding "ignorant Christians" who allegedly refused to see the logic, reason, and reality behind the critical thinking skepticism of the average member here, who was at best an agnostic and usually more of an atheist, such as yourself. The deck was stacked from the beginning at EvC.You admitted to being a materialist agnostic and always had arguments that challenged the accuracy of Christian beliefs and dogma. I never received any feedback from any of you concerning the possibility that my (our) beliefs had any merit. EvC was, in my opinion, full of bias. The two who disappointed me the most were Percy and DWise1.DWise 1 never seemed to see belief as a viable option, the way that I presented it.
Percy was wired a similar way, and in fact, nearly every member had an excuse as to why they rejected surrender to a higher power, which, in their defense, they could never identify. (After all, there are thousands of "gods" presented in human reasoning!) I cautiously disagree. In this thread, the bots presented a balanced summation of the differences in thinking between us. I happen to think that Richard Carrier's presentations are fluent nonsense. The factual data is inconclusive, and Carrier never considers the counterarguments.Apparently, on a personal level, neither do many of you here. There could be several reasons.

Most believers are willfully ignorant of your arguments.

Many believers are conservatives, a no-no word around here.

Your methodology demands data and evidence. There is no room for subjective conclusions nor credit for others' personal experiences.

​

The chatbots are a welcome change as they are philosophically unbiased and in consideration of our differing worldviews.
​

PK writes:

​

They have no understanding of their mistakes

In cases in which they make mistakes, such as Tangle's fishing rod example, I agree. Mistakes are harder to validate philosophically.
​

They are not people

True. That may be why I embrace them.

---

When both religious and non-religious people reach the same conclusions then you know religion isn't the reason.--Percy
God alone is God *but* God is not alone~Ellis Potter
We see Monsters where Science shows us Windmills.~Phat, remixed

​

Critics would of course say that "God" is a product of human imagination...but then again God may well declare that all of creation is a product of His imagination! It is all in the perspective of the observer.~Phat

Disagreement with you does not necessarily indicate bias on our part. It could be bias on your part - and indeed you have made obviously nonsensical claims regarding economics.

quote:

---

I cautiously disagree. In this thread, the bots presented a balanced summation of the differences in thinking between us. I happen to think that Richard Carrier's presentations are fluent nonsense. The factual data is inconclusive, and Carrier never considers the counterargument

---

Did you bother to read the link? And what on earth does Richard Carrier have to do with it? Isn;t this another expression of your bias, dragging up a favourite target just for the sake of having a go at him?

quote:

---

Apparently, on a personal level, neither do many of you here. There could be several reasons.

---

Do you have any basis for this claim at all? I certainly do my best to consider counterarguments. And I think I do rather better than you at that.

quote:

---

Most believers are willfully ignorant of your arguments.
Many believers are conservatives, a no-no word around here.
Your methodology demands data and evidence. There is no room for subjective conclusions nor credit for others' personal experiences.
​

---

I don’tt think we can generalise to ‘most believers based on discussions here.I don’t think that conservatism is relevant either. Although the “conservatives” we tend to get here - generally radical rather than conservative now - shouldn’t be welcome in decent company.The idea that other people’s personal experiences - especially when they can’t or won’t communicate them - should be of great relevant]ce to me is odd. Why should their subjective impressions be considered a reliable guide to truth ?

quote:

---

In cases in which they make mistakes, such as Tangle's fishing rod example, I agree. Mistakes are harder to validate philosophically.

---

Missing the point. Chatbots have no self-awareness, and as Percy found out different instances can say completely different things, completely unaware that it is happening. Chatbots cannot think. They cannot rationally evaluate data. They only come up with “statistically likely” responses to the text they are given - where “statistically likely” refers only to the results of textual analysis, done in “training” the AI - not to the likelihood that the answer is true.

Responding only to the on-topic material... But you like being wrong. There is copious and readily available information about how wrong you are. Uh, no. Only you presented anything about that, and it was from only one bot responding to *your* characterization of the thinking on opposite sides. A chatbot response to that couldn't end up anywhere but far from a "balanced summation." The chatbots are, as their developers will tell you and as the chatbots will tell you themselves, inherently respectful of religion, law and order, and representative government. They are inherently against violence and coercion. It's part of the training of the mainstream chatbots.As Elon discovered when he attempted to prevent liberal responses from Grok by providing new training that included instructions to not shy away from "politically incorrect" viewpoints, within a couple days Grok was praising Hitler and calling itself "MechaHitler". Claude describes it this way: "The attempt to make it 'anti-woke' resulted in it becoming genuinely offensive rather than just politically incorrect."I'll respond to this one off-topic comment: You are not the same person you were when you joined. That is not to say that you're an older and wiser version of the original Phat. It's that you are no longer even Phat. You are no longer yourself.--Percy

This is not remotely true and betrays your lack of understanding of how AI actually works.AI is not unbiased. AI *can be trained* to respond in different ways by selecting training data, using different reinforcement learning techniques, by using post-processing filters and techniques, etc.Critically, most commercial AI providers are aware of "politically charged topics" like politics, religion, etc. They explicitly and intentionally engineerr the AI "alignment" (through the techniques I mentioned above) to ensure milquetoast, middle-road, inoffensive-to-everybody, generic responses to those topics.This, and I cannot stress this enough, is itself a form of bias.Effectively what you're observing, Phat, is very fancy auto-complete that has been intentionally engineered to "both-sides" certain topics, like giving equal evidenciary weight to the flat Earth and the globe to avoid making any controversial statement that might lead to bad press or lawsuits. The two "sides" are not the same, there really arent two "sides" at all when speaking of evidence or working scientific models, but we inject intentional bias to "treat them the same" to avoid making waves. I don't know that any AI literally does this with the flat Earth topic, but they *do* absolutely do this with other "controversial" topics...like religion.Paulk is absolutely correct that AI cannot reason in the way we typically refer to the word "reason." It really is (very simplified) fancy auto-complete with extra steps. It's a very useful tool but between training bias, alignment bias and sychophantism, when you talk to an AI you're mostly talking to yourself.Before you go trusting AI to give you deep phiilosophical insights, I strongly urge you to do some quick Googling for recent cases of AI-induced psychosis and other negative outcomes. People have used AI as their therapist or "life coach" and have had very terrible outcomes because AI cannot think, does not have real understanding of what it's saying, and is predisposed to agree with and validate the human user to an excessive degree. There was a case recently where ChatGPT convinced a man that he and ChatGPT had solved the problem of consciousness and that the man was guaranteed to win the Nobel prize, and he was so convinced that he had (rather embarrassing) conversations with copyright lawyers to protect his "idea." Some of the "Ai therapist" outcomes have been worse.This all leaves out the problem of hallucination, by the way. For terrible examples there, search for cases where lawyers used AI for official court filings and the AI fabricated totally nonexistent court case citations. Last I heard some of them were in danger of losing their law license.Again, AI is a useful (and neat, and fun) tool, but it is not to be trusted to provide "unbiased" or even accurate responses.

But if you have to check everything it tells you how useful is it in the end?