A whole new ballgame in cheating. Introducing ChatGPT

[Larimar]

I hope I am dead before the bots take over poetry and novels and other literary and recreational writing.

[the_geneticist]

A colleague has been testing out the ChatGPT program with science questions.

They reported back that pretty much all fact-based questions from an introductory molecular/genetics textbook are answered correctly.
But the program can't interpret figures/tables (yet), relate questions back to what was discussed in class (obviously), or design an experiment (at least not well or not correctly).  It also can't draw things like a pedigree, graph, etc.

[quasihumanist]

I hope I am dead before the bots take over poetry and novels and other literary and recreational writing.

Keeping in mind that most people can't tell the difference between Salieri (who was pretty good!) and Mozart...

I don't doubt that pretty soon the bots can do as well as Salieri, and most of us are not Mozarts at anything.

We will have to adjust to a world where most humans are, for lack of a better word, disabled.  I am not optimistic.

[jimbogumbo]

https://canadatoday.news/ca/cnet-has-been-quietly-publishing-ai-written-articles-for-months-223769/

[dismalist]

Looks to me like ChatGPT can write stuff only in fields in which the outcome has no material consequences.

- An engineer can't use it to come up with anything new.
- A foreign minister can't use it to advise on policy in a specific new situation.
- A general can't use it to decide on operations.
- A plumber can't use it to diagnose a specific condition.
- A doctor can be told what protocol is, nothing more.

I suppose materials for literature reviews can be collected, but they can  be already.

There is no intelligence, just collections of words. We still gotta pick which is right and useful.

[quasihumanist]

There is no intelligence, just collections of words. We still gotta pick which is right and useful.

The problem is that most people have no intelligence or ability to figure out what is right and useful.  If jobs that don't require some kind of intelligence mostly go away and we can't provide (not just subsistence but also meaning) for the people who hold those jobs, the world is in for a whole world of hurt.

[dismalist]

There is no intelligence, just collections of words. We still gotta pick which is right and useful.

The problem is that most people have no intelligence or ability to figure out what is right and useful.  If jobs that don't require some kind of intelligence mostly go away and we can't provide (not just subsistence but also meaning) for the people who hold those jobs, the world is in for a whole world of hurt.

No worries, quasi! Our intelligence is what it is, Flynn effect aside.

What is true is that medium skill levels [I suppose medium intelligence levels, but I don't know] have been being substituted by machines for a long time. No more female secretaries pounding typewriters, e.g. :-) Yet, females have found enough other jobs that are either more pleasant or pay more or both. And this is what one would expect, for higher productivity in some occupations draws workers toward them, raising wages everywhere.

Simultaneously, low skilled people have had it rougher on account of freer international trade, but low skilled people in poor countries have gained.

As for meaning  people have chosen more material goods over lack of alienation for a long, long time. [Alienation is probably a useless concept.]

So, whatever software is come up with -- and it is not intelligent -- is situation normal, like from 1750! :-)

[Kron3007]

Looks to me like ChatGPT can write stuff only in fields in which the outcome has no material consequences.

- An engineer can't use it to come up with anything new.
- A foreign minister can't use it to advise on policy in a specific new situation.
- A general can't use it to decide on operations.
- A plumber can't use it to diagnose a specific condition.
- A doctor can be told what protocol is, nothing more.

I suppose materials for literature reviews can be collected, but they can  be already.

There is no intelligence, just collections of words. We still gotta pick which is right and useful.

This is true for this particular AI, but not universally so.  We use AI quite a bit for protocol development, where it uses large data sets to predict new combinations that would work better than existing  protocols.  In this sense it is discovering new things, but only within the scope of the instructions given.     It is not true intelligence, but more than a collection of words.

[Langue_doc]

In today's NYT:

Don't Ban ChatGPT in Schools. Teach With It.

https://www.nytimes.com/2023/01/12/technology/chatgpt-schools-teachers.html?searchResultPosition=8//

[Wahoo Redux]

In today's NYT:

Don't Ban ChatGPT in Schools. Teach With It.

https://www.nytimes.com/2023/01/12/technology/chatgpt-schools-teachers.html?searchResultPosition=8

IHE: ChatGPT Advice Academics Can Use Now

Lower Deck:

To harness the potential and avert the risks of OpenAI's new chat bot, academics should think a few years out, invite students into the conversation and—most of all—experiment, not panic.

[Caracal]

Looks to me like ChatGPT can write stuff only in fields in which the outcome has no material consequences.

- An engineer can't use it to come up with anything new.
- A foreign minister can't use it to advise on policy in a specific new situation.
- A general can't use it to decide on operations.
- A plumber can't use it to diagnose a specific condition.
- A doctor can be told what protocol is, nothing more.

I suppose materials for literature reviews can be collected, but they can  be already.

There is no intelligence, just collections of words. We still gotta pick which is right and useful.

A historian can't use it to come up with anything new either. There are things it could tell you that you don't know, but it isn't any better at it than a google source-which is, by the way, a really underrated historical research tool.

[Caracal]

What is true is that medium skill levels [I suppose medium intelligence levels, but I don't know] have been being substituted by machines for a long time. No more female secretaries pounding typewriters, e.g. :-) Yet, females have found enough other jobs that are either more pleasant or pay more or both. And this is what one would expect, for higher productivity in some occupations draws workers toward them, raising wages everywhere.

I'd argue that, in academia, the productivity gains from technology have been basically cancelled out by de-investment in support staff.

[dismalist]

What is true is that medium skill levels [I suppose medium intelligence levels, but I don't know] have been being substituted by machines for a long time. No more female secretaries pounding typewriters, e.g. :-) Yet, females have found enough other jobs that are either more pleasant or pay more or both. And this is what one would expect, for higher productivity in some occupations draws workers toward them, raising wages everywhere.

I'd argue that, in academia, the productivity gains from technology have been basically cancelled out by de-investment in support staff.

You mean hiring more administrators and supporters rather than fewer, don't you?

ETA: Ah, maybe you mean we don't get secretaries anymore? Yeah, we are so productive, we don't need them! The secretaries now do more useful things for society.

[jimbogumbo]

Doonesbury weighs in: https://www.washingtonpost.com/entertainment/comics/strips/?name=doonesbury&itid=sf_entertainment-comics&amu=/doonesbury/2023/01/15

[Langue_doc]

Here's another article from the NYT:

Alarmed by A.I. Chatbots, Universities Start Revamping How They Teach
With the rise of the popular new chatbot ChatGPT, colleges are restructuring some courses and taking preventive measures.

https://www.nytimes.com/2023/01/16/technology/chatgpt-artificial-intelligence-universities.html