Artificial Intelligence and Public Comment

I don't know...
Which is a fair stance. But still fairly naïve.
But see what it's doing now is just doing key word searches and copy pasting within the confines of a grammar program. It's just aggregating.

Anything in that category sure... but I guess I'm saying can it do art.

Not replication, sure it can copy Picasso, but can it be Picasso.
Autumn wind blows strong,
Elk hunt begins, silent tread,
Nature's symphony.


That's not the best haiku, but I've read worse ones from "artists"
 
Which is a fair stance. But still fairly naïve.

Autumn wind blows strong,
Elk hunt begins, silent tread,
Nature's symphony.


That's not the best haiku, but I've read worse ones from "artists"
Haiku's are an algorithm can it invent a Haiku? Can it want to?

It's just math at it's current level it's not being creative.

So back to @Nameless Range it can certain take folks current arguments and rework the grammar but it isn't going to analyze all the data and come up with a novel approach that will provide a middle ground for both sides.

Great for spam though...
 
I mean I know how competition works. And technically I'm in charge of the dept so it will ultimately be up to me whether it is used or not and how much. (I told him I'll set up some experiments and run my writing against ChatGPT and see how each performs and we'll go from there.)

My personal hangups center less on concerns for my own career viability, though I do have those concerns. Bigger concern is that I can't shake the feeling that asking computers to create prose, poetry, visual art and music is wrong and dark in some way that's hard to articulate.

TL;DR: I understand your argument in a utilitarian sense but I'm not sure that it tells the entire story. Hope my tone didn't come across as too snarky, btw.
Agree completely. But Capitalism pushes us toward the utilitarian pretty quickly. We have found it's not a choice of going back once something is created. It's not all that new. Computers have been writing summaries of companies quarterly earnings for a decade now. This just edges it forward a little more. Marginal jobs that require substantial writing, once thought only a human could do, will disappear quickly.
 
It can certain take folks current arguments and rework the grammar but it isn't going to analyze all the data and come up with a novel approach that will provide a middle ground for both sides.

...

It’s not clear to me that this is any different than what brains do.

I would wager that sooner than later it will be capable of generating that which is “novel” in a way that will be indiscernible from the way humans create novelty.

What’s happening on the backend may or may not really matter. What will matter is whether or not anyone can tell a difference.

Some would say Creativity itself is just a complex algorithm.Though you are right that this iteration is not that
 
Last edited:
Haiku's are an algorithm can it invent a Haiku? Can it want to?

It's just math at it's current level it's not being creative.

So back to @Nameless Range it can certain take folks current arguments and rework the grammar but it isn't going to analyze all the data and come up with a novel approach that will provide a middle ground for both sides.

Great for spam though...
Sorry, I can't drop this just yet. I found this interaction interesting. I asked the following.
1673631539790.png
What then happened was quite interesting. It drafted up a very nice essay describing both the various types, examples, and styles of art that can and is created by AI, but then provided a counter argument that while all that was true, it didn't have the cultural experiences that humans have to create truly unique art. It then concluded with "Therefore," and sat thinking on that for 15-20 seconds before throwing the above error, which I think was because it couldn't conclude that AI can create art.

All of that is really fascinating. But both you and it are talking about current limitations.

I see zero facts to support that will remain true in the future.
 
I think we quickly got to the idea that there are two parts - the form (the writing) and the content (the message). In many cases (the point of the OP), the content is an opinion. What if the content was factual? So I asked

Screenshot 2023-01-13 at 9.51.35 AM.png
It doesn't take much of an imagination to see that eventually our problem will become the Chatbot's problem - we get so much information, both real and fake, we can't tell the difference. How does it know the data source is legit? In addition, if Chatbot V1 has a limitation it won't write the letter I asked for (because of those little annoying things like facts) someone else could create Chatbot X that will not have this limitation. Or maybe someone floods a source with inaccurate info (or creates a bot to do it). We start the spiral down the rabbit hole...

I think it is a useful tool to help cut out the mundane task of creating the form, but it doesn't abdicate the reader from reading it and making their own conclusions. But I could say the same thing about Facebook posts...
 
Imagine being me, interviewing for a really good job this past Tuesday, one in which my writing will be "at least 50% of your responsibilities" and where my MA in English is a big part of why I'm able to argue for higher comp than the person I'm replacing. CEO says "We love and trust your writing ability and look forward to investing in it as we bring you onboard." As the meeting ended, he had one more question: "Have you heard of ChatGPT? Is there a way you can work that into your role here to save yourself time in your writing process?"

Yeah, I slaved through six years of school and poured my guts into becoming a writer, just so I can punch prompts in for a blank blank robot to spit back some tone-less soulless sentences we can pass off as me.

"John Henry, John Henry..."
Or you can sit at home and collect your universal basic income like most will. Our future is bright😂
 
I closed out the chat, but asked if AI could damage the envrionment. It said yes it can damage the environment or help it. I asked why have AI if has the potential to damage the environment. haha
I got this. One of the things I have found is if it denies you, you can ask the question in a different way.
View attachment 260550
Interesting. I asked the question 5 different ways, got the same response, until I made it only about jobs. Even then the AI bot had to toss in a comment at the end. haha I have a feeling that the "data" set used to generate the responses is so far one sided that it doesn't even look at the "other" data.

1673631451508.png

I asked a bunch of other questions (why I spent time arguing with a bot, that can just spit out garbage, I don't know), and the responses where all based on comments I've seen in response to the EIS or generated by NGOs against the project. Generalities and no specifics. The mine "could damage the environment:, well how? The mine has potential to impact the fish... How many? The responses are damn near verbatim of the Nature Conservancy, Save Bristol Bay, Earth Justice, et al websites. Speak in generalities with no specifics. Generalities = formulate opinion.

This one was my favorite. "relatively small" at 0.001%, but it's significant? It can't discern between magnitudes.

1673634223970.png
 
One final thought. The BOT is an environmental justice warrior, but has no issue making up fraudulent documents for you. Crazy times we live in.

Where it shines, and the reason so many are adopting it, is when you ask it to perform technical tasks. You can ask it to write code that is fairly complex and in multiple languages, and in seconds it will generate an output that not only works, but is commented well.

I know this only matters to a small subset of dorks on here, but it really is incredible.
 
One final thought. The BOT is an environmental justice warrior, but has no issue making up fraudulent documents for you. Crazy times we live in.
It's mostly a language processing tool (currently at least). If you want to find come grand conspiracy in it, that is your choice. I view it as a poor-man's secretary. It will write a readable, understandable message and eliminate spelling and grammatical errors. I'm still working on how much it "learns". As @Nameless Range points out, it apparently will write code in a fraction of the time. Again, both systematic processes historically done by people.

All I typed in was "write a letter showing the positive argument for Pebble Mine project". Maybe it knew you wanted to argue so it did that? ;)
 
Where it shines, and the reason so many are adopting it, is when you ask it to perform technical tasks. You can ask it to write code that is fairly complex and in multiple languages, and in seconds it will generate an output that not only works, but is commented well.

I know this only matters to a small subset of dorks on here, but it really is incredible.

So you're telling me it found stack overflow...
 
Back
Top