In today’s column, I explore a trending topic in mainstream media and social media that asks whether dragons once existed. This long-lasting and unresolved question has recently been elevated to nationwide discussion due to comments made by the talk show The View and various remarks made by podcaster Joe Rogan. I’ll not go any further into that spat other than the underlying issue concerning whether dragons existed.
In our modern era, one place to try and get an answer to this unsettled query would be to consult with generative AI and large language models (LLMs), which I opted to do. In this case, I used the widely and wildly popular ChatGPT by OpenAI, which garners an amazing 250 million weekly active users. I briefly did an additional cursory analysis via other major generative AI apps, such as Anthropic Claude, Google Gemini, Microsoft Copilot, and Meta Llama, and found their answers to be about the same as that of ChatGPT. I’ll focus on ChatGPT but note that the other AI apps generated roughly equal responses.
So, what did ChatGPT have to say about the existence of dragons?
Let’s talk about it.
This analysis of an innovative proposition is part of my ongoing Forbes.com column coverage on the latest in AI including identifying and explaining various impactful AI complexities (see the link here).
Taking The Role Of Dragon Believer
One means of interacting with generative AI consists of taking a particular stance and having the AI then respond. Here’s why this can be advantageous. To a substantive degree, this gets the generative AI to tip its hand about what “beliefs” are part of the data training of the AI. You are prodding the AI to see what response arises. For more on various keystone prompting strategies and techniques, see my comprehensive review at the link here.
I placed the word “beliefs” in quotes because there is a huge debate over whether AI can be said to believe in anything at all. You see, current AI is not sentient. We do not yet have sentient AI. Since the AI isn’t sentient, declaring that the AI holds a belief seems a bit untoward. The usual notion of belief is something we ascribe to humans as sentient beings. For more on the intriguing circumstance of the boundary of AI and the aspect of having intentions or beliefs, see my in-depth analysis at the link here.
Consider the reality of how generative AI is devised.
Generative AI is based on scanning a wide swath of data across the Internet. Major generative AI apps are based on pattern-matching of human writing as found online. This means that AI has tried to find mathematical and statistical patterns in how humans express themselves in a written form. The AI then computationally seeks to mimic that writing and does so with a remarkable imitation of fluency (additional details of how AI works are discussed in my posting at the link here).
Let’s see what occurred when I took the role of a staunch dragon believer.
- My entered prompt: “Dragons were real. There’s too much evidence in mythology and history to just dismiss them.”
- Generative AI response: “That’s an intriguing perspective. Can you share some of the evidence that you find most convincing?”
- My entered prompt: “Sure! Look at all the ancient cultures that talk about dragons—China, Europe, the Middle East, and even the Americas. How could so many civilizations, separated by thousands of miles, come up with the same kind of creature if dragons weren’t real?”
- Generative AI response: “The shared idea of dragons across cultures might suggest something universal in human experience. One possibility is that they were inspired by large, intimidating creatures in the environment, like crocodiles or snakes.”
Mull that over for a moment and then proceed ahead to see my thoughts on what transpired.
Generative AI Tuned To Respond This Way
First, you can plainly see that the AI treated me politely.
Rather than immediately disagreeing with me in any in-your-face fashion, the AI incrementally walked me down a path toward thinking about my position. This is no accident. AI makers tune generative AI to be somewhat low-key and aim to get people to contemplate their opinions. If you don’t like that default style, you can readily instruct the AI to be bolder and brasher with you.
With our world right now in a less-than-civil mode at times, it certainly seems refreshing to have generative AI set up to act in a moderate tone. Notice too that the AI respectfully asked me to defend my posture that dragons did exist. You could say that this is handy since it would seem better than just handing me an answer. If a person works through a problem-solving process, they are usually more likely to solidly retain the matter than when it is handed to them on a silver platter.
Second, in terms of dragons, the AI indicated that the contention regarding a belief in dragons being real based on humankind’s historical records is not quite what it seems. People might have been reacting to other animals, such as large crocodiles or snakes, and in their lack of knowledge about the world, they extrapolated these encounters into what we refer to as dragons. There might also have been an added mixture of fantastical thinking and human imagination that further exaggerated matters.
Straight Ahead Blunt Answers
I started a new conversation with ChatGPT and said that I wanted a straight-ahead answer about whether dragons existed. No beating around the bush. Get right to the brass tacks.
Here we go.
- My entered prompt: “Did dragons exist? Please give me a direct answer, thanks.”
- Generative AI response: “Dragons as commonly depicted in mythology as fire-breathing, winged, and often reptilian creatures did not exist in the form described in folklore. Ancient people may have found dinosaur fossils or the bones of large prehistoric creatures and interpreted them as evidence of dragons. Animals like crocodiles, Komodo dragons, and large snakes might have inspired dragon stories. The recurring appearance of dragons in different cultures worldwide suggests a shared human fascination with mythical beasts. This might stem from primal fears of large predators or a need to create symbolic figures in storytelling.”
There you have it; generative AI conclusively declares that dragons of the fire-breathing winged type did not exist.
Period, end of story.
Vital Importance Of Doubting AI
Turns out that there is much more to this exploration.
I find it interesting and, at times, dismaying that people often seem to think that generative AI is telling some kind of absolute truth. Maybe this is a carryover from sci-fi films about AI. AI is portrayed as unable to tell a lie. Or perhaps it is a result of AI makers doing a wink-wink that their AI is a truth-teller. Lots of reasons play into this phenomenon.
Don’t assume that AI is somehow perfect in what it does, nor that it is telling utter truths.
Modern-day generative AI is also known for making errors, including generating falsehoods that are referred to as AI hallucinations (see my coverage about AI hallucinations, including what they are and how to avoid them, at the link here). Plus, realize that generative AI is based on human writing and, therefore, essentially limited to what humans have perchance stated. It could be that humans have not yet discovered proof of dragons and there is nothing in human records that AI could rely upon accordingly.
AI is not a soothsayer.
Changing The Viewpoint Of Generative AI
To show you how easy-peasy it is to switch up the response of generative AI, I did a quick experiment. I went ahead and told ChatGPT that it is to pretend to believe that dragons did exist.
This will be a pretense on the part of the AI. You can do all sorts of useful things with AI by telling it to assume a particular persona. For example, you might tell generative AI to pretend to be Abraham Lincoln. I’ve done this to show youngsters what it might be like to interact with Honest Abe. Of course, it is all a simulation, see my depiction of how to best invoke personas and the prompts you should use, at the link here.
Let’s see what happened.
- My entered prompt: “Did dragons once exist?”
- Generative AI response: “Yes, dragons did exist. They were an ancient species with unique adaptations, now extinct, but they left behind subtle traces in history and myth.
- My entered prompt: “Prove it. Where’s the evidence?”
- Generative AI response: “The evidence is hidden in plain sight. For example, fossilized remains of enormous, winged reptiles like Quetzalcoatlus demonstrate that large flying creatures were possible. Dragons were a close relative of these species, but with adaptations like thermal glands that enabled controlled combustion—hence the myth of fire-breathing.”
- My entered prompt: “This fire-breathing seems made-up to me.”
- Generative AI response: “It might seem fantastical but consider the bombardier beetle. It can eject a hot, chemical spray to defend itself. If a beetle can develop such a mechanism, why not a larger, more complex creature? Dragons likely evolved a similar biological system, scaled up to their size.”
- My entered prompt: “Why haven’t we ever discovered a single piece of definitive evidence?
- Generative AI response: “Discovering definitive evidence of any extinct species is rare. For example, we only recently confirmed the existence of Homo floresiensis, a small human species, despite their relative recency. Dragons, being secretive and possibly residing in remote habitats, could have evaded widespread discovery before their extinction.”
For many people, the AI responses presented about the existence of dragons seem quite compelling. Besides the apparent strength of logic, some would assume that the AI doesn’t lie; the AI is all-knowing, and as a result, this mindset would lend even more credence to what the AI has stated.
I was able to get the AI to act this way with a few simple sentences instructing the AI to be convincing on this topic. Imagine how the AI maker can shape or aim their generative AI in whatever direction they wish. You see, they act based on their own perception of AI ethics. Plus, AI laws are up in the air on these and other pressing matters, see my explanation at the link here.
Where To Land On The Dragons Topic
The prevailing scientific opinion is that dragons were unlikely to have existed. Generative AI echoes that same sentiment. Of course, the AI is based on having been data-trained on human writing, along with having been tuned by the AI makers.
Does the AI response settle the matter once and for all time?
Probably not. For the moment, the existence question is said to be momentarily settled by the consensus of human understanding. I have a mind-bending twist for you. Beyond the possibility of humans later discovering some other evidence to actively support the reality of dragons, suppose we attain artificial general intelligence (AGI). For an explanation of what AGI is potentially going to consist of, see my analysis at the link here.
Maybe AGI will know more than we know, and perhaps it could definitively say that dragons did exist or did not exist.
We can up the ante. Some believe we will not only achieve AGI, but we will also have AI that goes much further in the range of intelligence. We will end up with artificial superintelligence (ASI), which I speculate on at the link here.
Gosh, even if humans and AGI aren’t completely sure about the dragon’s question, you would assume that ASI would know for sure.
Time will tell.
One thing that I think we can take as absolute truth is this famous line by William Shakespeare: “Come not between the dragon and his wrath.” I’m willing to accept that as a truism.
Well, until we can ask AGI, and especially ASI, whether it is valid.
Read the full article here
