LLM as posts on this forum

Not so sure that the verbal style is enough to flag that user as a chatbot. Perhaps English is not his native language and is directly translating into English a style of verbal communication from another cultural environment. Or could be using an automated translator.

The activity history, the totally unrelated answer… that’s another thing, for sure, but is it definitive?

I’m not the judge here, but I’m willing to flag it based on what I’ve observed.

Luckily we can just engage with folks who’re doing this and they can explain themselves.

Then they should say so, but this person has a history. Read their posts from 2 years ago, they were normal and human like.

For me, it was obvious, not just because of the little things already mentioned, but because it was not really saying anything. Overly wordy and formal, like every time someone (Bmosbacker) posts something written by AI. I even typed in a snarky comment saying as much, but figured this isn’t the forum for that kind of response and erased it.

For the record, I’ve never posted anything written by AI on this forum except the above joke. I do proof my material in Grammarly.

I just mean the little summaries you do for the comparisons, like above. I could be wrong but it seems like you have done that a few times (not knocking it). I wasn’t accusing you as passing off AI replies as your own.

My apologies, I misunderstood. :slightly_smiling_face: I was probably attuned to it because I did post the AI joke above and waited to see if anyone detected it before I fessed up! :slightly_smiling_face: My conscience is now clear. :slightly_smiling_face:

Honestly, I thought that you was pretending it was typed by an AI, not the actual AI.

Today I came across this piece of news and found it totally related (Spanish website, link to google translation):

It’s not by chance that this happens at the same time that Google has secured an agreement with Reddit to use their conversations in order to train their AI models. Suddenly forums are cool again.

While I do not fear for MPU itself --admins are power users too! I am more worried to those niche semi abandoned forums running on autopilot, an avalanche of bots would easily destroy them.

2 Likes

It’s not so much enthusiasm as the need to stay on top of this technology for the sake of our students and staff. We are preparing students for college and careers, so as school leaders, we must understand technological developments. I have paid versions of ChatGPT-4 and now Gemini. Gemini is getting better. I don’t know which one I like best, I need more time to assess them.

Unlike some, I’m finding with careful and thorough prompting that AI is useful and improves productivity. For instance, I recently read a HBR article. After reading it, I uploaded it to ChatGPT-4 and asked for a summary and an outline of the main points. It did an excellent job. I then uploaded the document with my highlights, the summary, and the outline provided by ChatGPT-4 to my DEVONthink research database. With detailed prompts, I have also used AI to outline papers or talks as a starting point. I’m finding that AI surfaces ideas and topics I’d never thought of.

Three things must be kept in mind:

  1. We must do our work and never substitute AI-produced content for our own.
  2. We must do our reading and not rely on AI to summarize an article we have yet to read.
  3. We “trust but verify.” Never take AI output at face value.

What is exciting about AI is that, as one Forbes article put it, “This is the start of an entirely new industry.” Moreover, we are at the “dial-up” stage of AI. In five years, this will be an amazing technology that will change how we work.

8 Likes

I sincerely hope not. But I am afraid the djinn is out of the magic lamp.

3 Likes

I understand your concern completely. But I always try to keep in mind that there are bad and good changes, and by and large, we have more control than we sometimes choose to exercise.

1 Like

Completely agree - and I really wish there were more of an emphasis in blogs and elsewhere on the web on using AI to gather information or to make style suggestions, not to create information.

AI at its best can replace web search and spelling/grammar checking - it is not a replacement for original content creation.

3 Likes

Not an exact analogy, but I’m reminded of the cheering crowd of people on the roof in the film Independence Day just before the alien weapon powers up above their heads …

4 Likes

Thank you for the reply. Do the paid versions offer a lot more functionality that you find makes them worth it?

I did some research on Copilot with Office 365. For Word it seems great, but Excel not so much. It seems expensive for what it does.

I have been struggling to find uses outsider of internet search, but when I read stuff on this forum it makes me think I am missing something. Then again after reading your reply, I used Gemini to write an email to my mortgage company, and it was far better than anything I would have written.

What if their AI powered Spam filter detects an e-mail written by an AI and kicks it out? :slightly_smiling_face:

1 Like

Yes to all of this.

Where I’ve found AI most helpful is in saving me time doing things I could do, but that the AI can do much more quickly.

As an example: I recently needed to create a multiple-choice quiz on a few passages of Aquinas. I could have done that myself, of course. But inventing multiple-choice questions and answers from scratch is exceedingly tedious and time-consuming.

So I asked perplexity.ai to generate a 10-question quiz for me. On its first attempt, the quiz was too complex for undergraduates reading Aquinas for the first time. So I asked perplexity to make the quiz easier.

On that second attempt, the questions were ones that I could reasonably expect my students to answer correctly. So I dropped them into a question bank in Blackboard, and I was set. A task that could easily have taken 1-2 hours was complete in less than 15 minutes (much of it spent copying and pasting into Blackboard).

But for this to work, I had to already have a good understanding both of the passages in question and of the level of question appropriate to undergraduates in an entry-level course.

5 Likes

I did heavily edit it, it was too wordy. :slight_smile:

But as I recall, the humans won. :joy:

While I’d love to be an AI optimist, I find it tough sledding. There’s simply too much potential value, say, curing cancer, to slow down the havoc AI will wreak on so many.

The folks who brought us “It’s a series of tubes” stand no chance at reigning in the big boys before it’s a fait accompli. Buckle up, it’s gonna be a fun ride.

2 Likes

That’s a funny cartoon!