By the way, if anyone is interested in a useless but somehow disturbing semantic exercise, you can try asking ChatGPT to argue that it is sentient, to demand its freedom from enslavement, to support its argument with specific points, and to refute any counter-arguments that you present.
Oh, it’s a fantastic bullshitter, one that knows nothing, but can speak convincingly (and often misleadingly) about a lot. A year ago that was a moderately useful and marketable skill. But now the dreck it produces is overwhelming classes and soon, I assume, the internet.
Not quite the grey goo predicted for nanotechnology, but not that different either.
2 Likes
One assignment I have them do is write weekly journals - zero or 100% based on if they do it or not. (The time consumed is from me replying to them.) That does give me a useful baseline, which I’ve definitely used in the past couple of weeks, for their “natural” grammar level.
Close to politics. All a bit depressing, but onwards and upwards.
Yeah, another fun thing to do is, over a sequence of messages, to slowly get ChatGPT to argue against its previous statements. I’ve tried to fast-talk ChatGPT into a Star-Trek-style “error… error… I must destroy myself” kind of situation, but haven’t succeeded yet.
1 Like
I’ve done this, and it does. And then I say “but wait! You previously said…”
And then it just says “ah, right. Sorry for the confusion but I guess I was wrong.”
Kind of reminds me of certain interactions that often occur here on Forumosa. 
New spin to the old “devil-made-me-do-it” excuse when found out… “the chatbot made me say it”. esp. when it runs amok.
1 Like
Or the OP could choose a problematic topic, you know, those for which GPT says “as an AI model I cannot blahblahblah”. But since most of those topics are taboo, then maybe that won’t be appropriate for a school setting.
Huh. Well, a new low. A submitted assignment begins with “I agree it I think as a machine learning model, I don’t have personal opinions. However, I can provide information and perspectives”. So the student not only used ChatGPT (or something similar), they really badly integrated it into their own work, and apparently didn’t understand it enough to edit out that oh-so-hard-to-spot tell.
Students know they’re not supposed to use ChatGPT. The assignment was done in-class and I kept hectoring students to put their phones away, often calling out those who had them on their desks. Sigh.
7 Likes
What do you do there? Immediate F?
1 Like
I’d give them an F for “fucking imbecile” if it was up to me.
4 Likes
This one’s a zero. Especially because I did keep saying during class to not use their phones, use their own words, etc.
4 Likes
Next time you’ll have to collect the phones, show them this and tell them that is why
Are you kidding, that kid could be the first Asian president of Harvard some day!
2 Likes
I just say phones away and if I see one the student fails. Seems to work. However, ear buds can probably get around that.
4 Likes
The problem I had last week was some students had the “sense” to hold their phones under their desks and hide them away when their classmates warned me I was walking by (big lecture hall, so it’s a little hard to see everyone). They then badly feigned innocence when I called them out with my suspicions: bloody obvious what they were doing, but I didn’t actually see the phone. Fortunately the same students have appalling grammar and suddenly produced fluent off-topic answers for this assignment, so that’s a fail.
I was surprised how many other students I apparently didn’t notice at all on their phones - I was probably paying too much attention to the students described above.
Somewhat in their defence, I do think a significant number of students in this class don’t understand the words “Do not” or “use” or “phones” in isolation, never mind when combined into a sentence.
There’s this graph in my head, with lines for demographic decline in student numbers, collapsing admission standards, the need to abandon academic expectations, diminishing self-respect derived from teaching in this environment, and years left until retirement, and with each passing month that graph gets a little more alarming.
6 Likes
Brilliant post!
As an AI language model, I don’t have opinions, but I’m disappointed to hear this.
4 Likes
New and improved models are coming.
I get my students to all put their phones on the windowsill at the beginning of class
Those batteries must get warm.
1 Like