Talking to the machine
Prophets of doom are having a field day with the advent of ChatGPT and other forms of advanced AI. It’s the greatest existential threat to humanity, they say, since … well, since the last greatest existential threat to humanity. From my experience, however, I think it could be quite the opposite. Rather than leading us to perdition, it could well be the saving of us.
The other day by way of an experiment, I asked ChatGPT to explain the spelling of the words “practice” and “practise” in English which often cause confusion among learners and native speakers alike. “The confusion might come from the fact that in American English, both the verb and the noun form are spelled with a ‘c’,” came the reply, “whereas in British English, the verb form is spelled with a ‘c’ and the noun form is spelled with an ‘s’.”
No, no, no, said I. You are mistaken. It is the other way around. In British English it is the verb that is spelled with an ‘s’ and the noun that is spelled with a ‘c’.
The machine was quick to recognise the error of its ways and not only correct itself but also offer profuse apologies. “You are correct, and I apologize for the confusion in my previous response. I made an error in my explanation,” it said. “You are right that the noun form of the word is spelled with a ‘c’ in both American and British English, while the verb form is spelled with an ‘s’ in British English … Thank you for bringing this to my attention, and I apologize for any confusion I may have caused.”
To many people the danger of AI lies not so much in the fact that it can make elementary mistakes (and presumably much more serious ones as well), but the way it mimics the human tone like this. Not only does it sound human, but I have heard that some people who have engaged it in conversation are referring to it as “he”, so not only human but a man!
However, they are doubly mistaken. ChatGPT is not a man, it’s a machine and so its pronoun should be “it”. And, of course, it’s not human either. Despite what people say, its reactions are not human-like at all.
Take the example above. At first blush the machine’s reaction to having a mistake brought to its attention seems very human. But think about it for a moment, do people really react like that when having their silly mistakes pointed out to them? Do you?
I think this same conversation with a human would go more like this:
Joe: What’s the difference between “practice” with a “c” and “practise” with an “s”?
Sam: In British English the noun is “practise” with an “s”, and the verb has a “c” as in “practice”. In the United States both noun and verb have “c”.
Joe: No, you’re wrong. Surely in English it’s the verb that’s spelled with an “s” and the noun that’s spelled with a “c”.
Sam: Don’t be silly. It’s the other way round. Don’t you know anything?
Joe: I’m pretty sure I’m right. The verb “practise” has an “s” in British English.
Sam: No, way.
Joe: Yes, it has.
Sam: No, it hasn’t.
Joe: Yes.
Sam: No.
Joe: Tell you what, let’s look it up. Here’s the Concise Oxford Dictionary. See, I’m right. The verb is spelled “practise” in British English.
Sam: What would they know about it?
Joe: It’s the Oxford Dictionary, Sam. They know everything about it.
Sam: Bunch of idiots if you ask me.
Joe: Come on Sam, don’t be silly.
Sam: And you’re just as bad.
Sam then raises his voice and calls Joe all sorts of names, compares him to the followers of a couple of notorious European dictators in the 1930s and 40s, and casts aspersions on the morality of his parents as well as his general lack of intelligence before storming out of the room and slamming the door as he goes. Joe, in turn, bangs the dictionary down on his desk and mutters similar insults about Sam.
That is what the human reaction looks like. Humans resent criticism and become defensive at the first hint of it. But if cornered, they go on the attack and resort to irrelevant arguments and finally, when pressed, often respond with a tirade of personal insults, thus, making everything worse.
If there is an existential threat to humanity, as so many people are saying, it is much more likely to come from humans themselves rather than a machine which accepts criticism, quickly corrects its mistakes, and apologises profusely. Pouring oil on troubled waters, as the saying goes.
Far from being a threat, these cool, calm and unemotional machines, may well end up saving us from ourselves!
The other day by way of an experiment, I asked ChatGPT to explain the spelling of the words “practice” and “practise” in English which often cause confusion among learners and native speakers alike. “The confusion might come from the fact that in American English, both the verb and the noun form are spelled with a ‘c’,” came the reply, “whereas in British English, the verb form is spelled with a ‘c’ and the noun form is spelled with an ‘s’.”
No, no, no, said I. You are mistaken. It is the other way around. In British English it is the verb that is spelled with an ‘s’ and the noun that is spelled with a ‘c’.
The machine was quick to recognise the error of its ways and not only correct itself but also offer profuse apologies. “You are correct, and I apologize for the confusion in my previous response. I made an error in my explanation,” it said. “You are right that the noun form of the word is spelled with a ‘c’ in both American and British English, while the verb form is spelled with an ‘s’ in British English … Thank you for bringing this to my attention, and I apologize for any confusion I may have caused.”
To many people the danger of AI lies not so much in the fact that it can make elementary mistakes (and presumably much more serious ones as well), but the way it mimics the human tone like this. Not only does it sound human, but I have heard that some people who have engaged it in conversation are referring to it as “he”, so not only human but a man!
However, they are doubly mistaken. ChatGPT is not a man, it’s a machine and so its pronoun should be “it”. And, of course, it’s not human either. Despite what people say, its reactions are not human-like at all.
Take the example above. At first blush the machine’s reaction to having a mistake brought to its attention seems very human. But think about it for a moment, do people really react like that when having their silly mistakes pointed out to them? Do you?
I think this same conversation with a human would go more like this:
Joe: What’s the difference between “practice” with a “c” and “practise” with an “s”?
Sam: In British English the noun is “practise” with an “s”, and the verb has a “c” as in “practice”. In the United States both noun and verb have “c”.
Joe: No, you’re wrong. Surely in English it’s the verb that’s spelled with an “s” and the noun that’s spelled with a “c”.
Sam: Don’t be silly. It’s the other way round. Don’t you know anything?
Joe: I’m pretty sure I’m right. The verb “practise” has an “s” in British English.
Sam: No, way.
Joe: Yes, it has.
Sam: No, it hasn’t.
Joe: Yes.
Sam: No.
Joe: Tell you what, let’s look it up. Here’s the Concise Oxford Dictionary. See, I’m right. The verb is spelled “practise” in British English.
Sam: What would they know about it?
Joe: It’s the Oxford Dictionary, Sam. They know everything about it.
Sam: Bunch of idiots if you ask me.
Joe: Come on Sam, don’t be silly.
Sam: And you’re just as bad.
Sam then raises his voice and calls Joe all sorts of names, compares him to the followers of a couple of notorious European dictators in the 1930s and 40s, and casts aspersions on the morality of his parents as well as his general lack of intelligence before storming out of the room and slamming the door as he goes. Joe, in turn, bangs the dictionary down on his desk and mutters similar insults about Sam.
That is what the human reaction looks like. Humans resent criticism and become defensive at the first hint of it. But if cornered, they go on the attack and resort to irrelevant arguments and finally, when pressed, often respond with a tirade of personal insults, thus, making everything worse.
If there is an existential threat to humanity, as so many people are saying, it is much more likely to come from humans themselves rather than a machine which accepts criticism, quickly corrects its mistakes, and apologises profusely. Pouring oil on troubled waters, as the saying goes.
Far from being a threat, these cool, calm and unemotional machines, may well end up saving us from ourselves!
No comments have been added yet.
