I think the issue is more macro than an AI generated piece of content. It's more the fact of the matter that an AI was deployed to communicate to students in the aftermath of a mass shooting at a university, rather than a human touch. Why was the author even considering using AI in the first place? You're communicating to a bunch of students who are scared shitless that a shooting could happen at their school, because they've grown up with school shootings being the norm. This is the first generation born with Columbine fresh in their parents' minds, and they had to come of age with the likes of Sandy Hook, Parkland, and Virginia Tech happening. It should sound obvious, but such communication requires a human touch.
I'm not of that generation, but it's my impression they don't care whether the empty words come from a person, a politician, or a greeting card company; and they certainly don't care if the empty words come from the brain of a university administrator or a website they visited. The words are all equally empty.
The only thing that matters is a cultural change that will lead to fewer mass shootings. Getting angry at university administrators for whatever lack of empathy they demonstrate is myopic. There's plenty enough reason for people to get angry about the way things are, but this is the wrong thing to be mad about.
> Why was the author even considering using AI in the first place?
Maybe because there aren't anymore words to be said about these things. They happen so frequently, that the expectation is words will be said after the fact, rather than action will be taken to prevent the next occurrence. So here we are, days after the shooting, mad at the words someone said after the shooting, rather than the fact there was a shooting.
Who here is still mad about the shooting? Who here is going to do something today to make the next shooting less likely?
>Who here is still mad about the shooting? Who here is going to do something today to make the next shooting less likely?
I'm still mad about Sandy Hook. That's in my home county. I'm still mad about Tucson - that was the local grocery store I went to during school. I'm still mad about Virginia Tech, I'm still mad about Vegas, and every other mass shooting. I'm still mad that it took Alex Jones nearly a decade to be convicted for the continued harassment and verbal abuse of people that had to deal with their 6 year olds get shot up by a psycho while sitting in 1st grade class.
Unfortunately, there's a powerful voting bloc both in DC and around the US that will vote to make sure nothing happens. Forget about enforcing the laws on the books - lots of police sheriffs around the nation have said they have no interest in doing so.
> But that’s my question. If the AI generated the exact same words a human touch person would have is disclosing the author necessary?
Honestly, the disclosure probably had nothing to do with the message itself. I'm sure it was all about institutional beliefs and policies regarding plagiarism. If some student took credit for an unattributed ChatGPT essay, they'd probably be pitilessly be given a zero at best.
However, including the disclaimer totally undermined the message (which was a pro-forma thing no one would probably have cared about otherwise). The message said "we care," but the disclaimer said "but not that much," and that would understandably offend people if noticed.
But it isn’t about the plagiarism. It’s confirmation that the institution only cares about empty platitudes while not really doing anything else to help their students. And if they’re not helpful, why does their job exist?
Empty platitudes are all they have to offer. What are they supposed to do? Lie and say it'll never happen again? Lie and say they have the power to prevent it? Or to effect meaningful change? Tell the truth that the students in fact are at risk and there's nothing anyone at the university can actually do to prevent it?
The futility of the whole situation has resulted in a culture of ceremony around the devastation as a way to cope. Now that people have violated the coping ceremony, we're mad at them. But why? Because we are powerful enough to at least scold them in an attempt to preserve the coping ceremony. But they're not the ones shooting people. They're not the ones directing policy that causes mass shootings.
Instead, we need to direct all that anger to changing the situation such that they never are put in a position of writing letters about shotting victims. Instead of smugly stating they are worthless people because their jobs can be replaced by a chatbot, we should question why it is the case that someone's job is to write condolence letters after shootings. Why aren't we mad that's an acceptable job function in the US? The solution students are looking for is not more empathy toward shooting victims, it's fewer shooting victims in the first place!
> But it isn’t about the plagiarism. It’s confirmation that the institution only cares about empty platitudes while not really doing anything else to help their students. And if they’re not helpful, why does their job exist?
You're kind of missing the point:
1. Institutional practices/beliefs about plagiarism will influence its behavior in other areas.
2. "Empty platitudes," are often necessary, as a social lubricant. People aren't robots. Sometimes you can't do anything, but also you can't say nothing.
3. I very much doubt the university administration only exists to send out these messages, so you can't reasonably question if their job should exist based on this narrow episode.
Think about the effect on the writer, in one case you need to sit down and face the blank page finding the right words to reassure, calm down and empathise with young adults who are scared and look up to you. You might even change as a person, just by the mere act of having to find those words and thinking about what those kids need and feel.
In AI generated text you can just ask for platitudes, copy paste, send and never have to confront anything. You can borderline become a worse person by becoming calloused at how easy it is to have to face a difficult situation
The words themselves are fairly unimportant in this whole situation, what people want out of the letter is to be heard, understood and have actual leadership. By having a glorified clippy write it for you, you are letting them down regardless of the quality of the text
So you're saying the problem is that the administrator didn't reflect enough on the situation when using AI to compose the the letter. But how do you know? Is the problem that the recipients of the letter aren't sure whether the author grew while composing it?
I am saying the point of even sending the letter is to demostrate leadership and empathy. Two things delegating the task to a pimpied up suggestion engine fails miserably at.
Yes. You can argue about the bytestrings being identical until the cows come home, but the vast majority of people believe that natural language text has invisible metadata tags denoting who wrote it and why. When the task is “communicating with people”, you have to respect their consensus reality.
Warning: cows coming home. Invisible meta tags? That sounds like a religion faith argument vs science! Something we can’t see and a string comparison == comes back true and yet somehow there is a difference? ASCII text is just numbers. If all the numbers are the same I don’t see where the invisible meta tags exist unless you mean like in a spiritual plane.
No, you’re right, the invisible metatags do not exist, not even in a spiritual plane. Nevertheless, most people believe they do exist, although they would not use the phrase “invisible metatags”.
To put the “respect their consensus reality” claim into a different format: every computer’s implementation of == will return true for the two strings, but most human brain implementations of == will return false - i.e. the difference isn’t in the string, it’s in the comparator function. The human brain implementation is obviously bugged, but the maintainer (whether God or evolution) isn’t accepting pull requests, so the bug is unlikely to ever be fixed. If you have to interface with the human implementation anyway, you simply have to account for this oddity. It’s like supporting Internet Explorer on your website.
By definition, the AI cannot and will never generate the exact same words as the PR person themselves. The community wanted the university spokesperson to respond. They did not, they outsourced the job.
And you might say "what's the difference???" and what you have to realize is, this is entirely emotional and that is the point. The community wants real empathy. They do not believe they are getting it, and you can't logic your way out of that.
I think that the way you get to the content is also important. If you are looking at a copy (good enough so you can't tell the difference) of painting by Rembrandt, you should be informed that it is a copy.
Same with words. Provenance for text will be a thing.
