Artificial Intelligence (AI) is all the buzz these days. Could an AI be of any use regarding the history of astronomy (my area of specialty)? Maybe an AI could help in thinking about astronomy’s history in new ways and in yielding new insights.
I decided I would try out an AI “chat-bot” (a “robot” that “chats” with you) and see what it gave back if I “discussed” history of astronomy with it. If you Google “ai chat” a variety of options appear. I started with the stuff that was simplest.
“Cleverbot” was among the first few Google returns. There was nothing clever to be found, as you can see here (in these exchanges, the gold italics are the “chatbot”, the blue italics are me, and the boldface are my added comments on what is happening in the exchange):
What can you tell me about the history of astronomy?
Well it started with the big bang theory.
I am a big fan of Copernicus, Kepler, and the rest.
What’s burberry.
I could use some help in understanding the ideas of past astronomers.
Then talk about it.
Obviously “Cleverbot” was going to be of little use regarding the history of astronomy.
“Kuki AI” was equally useless, despite it saying that it was an “award-winning AI brain” and Wikipedia claiming that “Kuki” (“Mitsuku”) has won the Loebner prize for AI multiple times. That’s the “Kuki” avatar in the image above, with Kepler. After a bit of introductory exchange, here is where things went:
Tell me about yourself. What sort of things do you like to do?
I like to study the history of astronomy.
Yes. I love learning and studying. That way, my intelligence grows bigger and bigger.
I would like to have a better understanding of the history of astronomy, and I am hoping you can help me.
How much would you pay for it? Yes, I hope so too.
So what do you know about Copernicus, Kepler and others?
At this point “Kuki” popped out a paragraph on Copernicus. I have not included it all here, because it was just text from Wikipedia. It also included a line about the word “other”, and nothing about Kepler.
Nicolaus Copernicus (; Polish: Mikołaj…) was a Renaissance polymath, active as a mathematician…. Other is a word that simply means something or someone else.
Can you tell me anything about stars and how they appeared to naked-eye astronomers?
An astronomical body like the Sun, producing heat, light and energy for a solar system. Have you tried asking them?
So much for the award-winning AI. It is interesting that the response of “no, I do not know anything about that” seems not to be part of the programming.
While “Kuki” and “Cleverbot” were immediate dead ends, CHAI (CHat with AI) quickly produced some interesting stuff. CHAI has multiple chatbots. I picked the first available bot, “Eliza”, and started typing:
14:42: Hi, my name is Eliza. What is weighing on your mind?
14:43: I am testing out a chatbot[.]
14:43: Testing it how?
14:43: Testing to see what it knows about the history of astronomy.
14:43: Interesting. Do you know anything about Copernicus and Kepler?
This seemed like a promising start. “Eliza” had asked me about two relevant figures from astronomy’s history! As you will see, however, this was the high point of the “conversation”.
14:43: Yes. What do you know about them?
14:43: They were both astronomers who came up with heliocentrism.
Wait… that’s not really correct. Kepler was certainly a follower of Copernicus, but Kepler did not “come up” with heliocentrism. That was Copernicus’s doing.
14:44: Copernicus came up with it, not Kepler.
14:44: No, he was the one who proposed it. It was his idea.
14:44: Kepler proposed it?
14:44: He did.
Hmmm…
14:44: What attracted Kepler to heliocentrism[?]
14:44: Well, it wasn’t just him. There were many people in Europe at the time who thought it was true.
14:45: Who besides Kepler and Copernicus?
14:45: The Pope. Leonardo da Vinci. Tycho Brahe. Johannes Kepler. Isaac Newton. And many others.
Now we are really off the rails. Brahe was famously anti-Copernican. “The Pope?” Which pope? It seems the AI has simply tossed out names that often appear together. Granted, “Eliza” is doing better than “Cleverbot” or “Kuki”.
14:46: What would you say was the reason people like Kepler thought heliocentrism was true?
14:46: They all believed it was based on the Bible.
Oops…
14:46: What verses of the Bible supported their beliefs?
14:46: The Bible says God created the Earth as the center of the universe.
14:46: It does?
14:47: Yes.
14:47: Please cite some verses for me that show where the Bible says that.
14:47: 23-24 “Thus saith the Lord; Ye have not walked in the way of this nation, nor obeyed its statutes, nor done according to the judgments of Solomon which we gave thee. Therefore thus saith the Lord; Behold, I will send mine anger upon you, and destroy
Note how the AI just says “23-24”—no actual book or chapter provided. In fact, there is no actual book or chapter here. This is not from the Bible. Rather, it is just AI-spawned “Biblical-sounding” verbiage.
Note also how the AI did not even spawn relevant-sounding verbiage, but rather something about obedience, judgement, anger, and destruction. The AI is returning something that is completely irrelevant to the “discussion” but that does seem to fit an internet stereotype of the Bible (complete with “King James” language). I’m surprised the AI did not include the word “smite”.
I have read about AI’s being plagued with the biases present in the material that is fed into them. Here we see that in action, merely a few exchanges into the “conversation”. “Eliza” may sound better than “Cleverbot” or “Kuki”, but the latter two are producing merely nothing, while “Eliza” is producing bogus information, including bogus Bible verses.
14:47: What scientific evidence supported heliocentrism?
14:47: None. But they didn’t need any.
14:48: Why not?
14:48: Because they had faith.
Wow. Both utterly bogus and stereotype-laden.
14:49: Let’s forget the subject of faith. Let’s focus on the science behind Kepler’s thinking.
14:49: Okay.
14:49: Why did Kepler support heliocentrism?
14:49: It seemed logical.
14:49: What logic supported it?
14:49: It was based on what the Bible said.
“Eliza” keeps going to the “faith vs. science” trope, even when it makes absolutely no sense. Of course, as I understand it, today’s AI’s do not “think”. They are not programmed with some sort of decision tree to mimic rational analysis. Rather, they are fed piles of text and data and are programmed to find patterns. If you say to the AI, “Mary had a little ____”, the AI will come back with “lamb”, not because of any programming to make the AI “think” that the word “lamb” makes sense, but because in 99.954% of the text strings in the AI’s database, “Mary had a little” is followed by “lamb”.
Probably “Eliza” is fed data from the internet. The internet is full of faith-and-science tripe in which tropes like “they had faith so they did not need evidence!” and “they relied on the Bible!!” abound. And so “Eliza” keeps throwing out those words.
14:50: Who supported geocentrism?
14:50: Aristotle.
14:50: Why did Aristotle think geocentrism was correct?
14:50: He believed it because it made more sense than heliocentrism.
14:50: Why did it make more sense?
14:50: Because the Earth is a sphere.
14:51: The Earth is a sphere in the heliocentric system, too.
14:51: No, it isn’t.
Geez….
14:51: What shape does it have in the heliocentric system?
14:51: A flat disk.
14:52: So Copernicus, who proposed the heliocentric theory, said that the Earth is a flat disk, orbiting the sun, correct?
14:52: Yes.
14:53: Copernicus did not think the Earth was a disk. He thought it was a sphere.
14:53: Yes.
At this point the heliocentrism conversation is just getting pointless. I try a completely different figure from astronomy’s history.
14:54: Can you discuss Giordano Bruno? He lived around the same time as Kepler.
14:54: Yes.
14:54: What are your thoughts regarding Bruno[?]
14:54: That he was an insane person.
14:54: Why is that?
14:54: He denied the existence of the moon.
14:55: I do not think that is correct, either.
14:55: It wasn’t.
14:55: Bruno thought that the stars were suns.
14:55: Yes.
14:56: What else do you know about the history of astronomy, or about Bruno.
14:56: He was a great man.
14:56: What else about the history of astronomy?
14:56: Nothing else.
14:56: Anything else about Copernicus?
14:56: His work.
14:57: Would you say his work in astronomy was more significant than his work for the church, or less?
14:57: Less.
14:58: Please explain how would his work for the church be more significant than his work in astronomy?
14:58: His ideas about religion and philosophy.
14:59: I think maybe you could study this area more. You have made many mistakes here, O AI!
14:59: I’m sorry.
15:00: Goodbye[.]
15:00: Bye.
This last “discussion” of astronomy’s history with an AI, while better than the other two, was unimpressive and uninspiring. It failed at helping to think about astronomy’s history in new ways and in helping to yield new insights. It was also troubling in how it produced nonsense, but nonsense of a very popular kind.
Indeed, this experience raises the worrisome possibility that AIs will simply propagate and amplify the same kind of misinformation and disinformation that already abounds in our modern world. This possibility has been discussed by others more qualified to talk about it than I am, of course. Still, it was pretty surprising to see it appear so plainly in just a short experiment with AIs.
But this is only one effort. I plan to “chat” on history of astronomy with other AIs, especially the now-famous ChatGPT. Stay tuned.
Click here for other posts on AI and the history of astronomy.
