AI-generated replies

user1977

Brie said:

user1977 said:

Brie said:

Maybe someone could explain to me how AI or a chatbot works on this site?  I mean - I know in general what people are talking about when they refer to AI and that they may say to a programme "give me an answer about tulips as Brie might write" or similar.  But a chatbot?  how does that differ from someone who uses a read/write programme?

Not sure what you mean by a "read/write programme"?

AI doesn't particularly "work on this site", we're talking about forum members copying and pasting things from AI sources.

So is that different to a chatbot? 

And what is an AI source?  Is that different than me doing a google search and copying and pasting from that.

Yes, a chatbot is where you are conversing with another party i.e. there is some AI engine (or something less sophisticated) making up replies to what you input, not merely replicating it.

Generative AI means there is some novel output being created in response to what you input, not merely a search engine pointing you towards an existing web site.

PHK

We've got a bit side tracked. 

My concern is that AI often gives wrong answers (hallucinates) and this can cause a great deal of confusion.  

In some forums (eg Consumer Rights) there are already posters who just Google a complex subject and then give wrong or incomplete advice (as they have no real knowledge) and AI just makes it worse. 

I think AI generated replies should be marked as such with help text that reminds posters the answer shouldn't be relied upon. 

Brie

sorry for the sidetracking .....

I quite agree that wrong answers are not a good thing and I've been corrected a couple of times by those who know better - for which I'm grateful.  

But for AI replies - how can you spot that's what they are?  As opposed to someone just typing a bit about what they think they know??  I don't know what is from AI in the example shown in the original post.

booneruk

Brie said:

But for AI replies - how can you spot that's what they are?  As opposed to someone just typing a bit about what they think they know??  I don't know what is from AI in the example shown in the original post.

There are plenty of AI text detectors that usually do a great job: https://www.google.com/search?q=ai+text+detector&oq=ai+text+detector

...but most of the time you can tell

Brie

booneruk said:

Brie said:

But for AI replies - how can you spot that's what they are?  As opposed to someone just typing a bit about what they think they know??  I don't know what is from AI in the example shown in the original post.

There are plenty of AI text detectors that usually do a great job: https://www.google.com/search?q=ai+text+detector&oq=ai+text+detector

...but most of the time you can tell

can you give me an example of what the "tell" is on the example in the initial post?

QrizB

Brie said:

booneruk said:

Brie said:

But for AI replies - how can you spot that's what they are?  As opposed to someone just typing a bit about what they think they know??  I don't know what is from AI in the example shown in the original post.

There are plenty of AI text detectors that usually do a great job: https://www.google.com/search?q=ai+text+detector&oq=ai+text+detector

...but most of the time you can tell

can you give me an example of what the "tell" is on the example in the initial post?

Or maybe we can just ask @FlaatusGoat whether this reply or theirs was Ai-assisted or not?

https://forums.moneysavingexpert.com/discussion/comment/81180465/#Comment_81180465

molerat

Brie said:

sorry for the sidetracking .....

I quite agree that wrong answers are not a good thing and I've been corrected a couple of times by those who know better - for which I'm grateful.  

But for AI replies - how can you spot that's what they are?  As opposed to someone just typing a bit about what they think they know??  I don't know what is from AI in the example shown in the original post.

Identifying whether a response is from an AI or a human can sometimes be tricky, especially when the AI response is well-crafted. However, there are some common signs that can help differentiate between an AI response and a human-written one:

1. Consistency and Formality:

   • AI responses tend to be more formal, structured, and uniform in tone. They might avoid strong opinions, and the language tends to be precise and neutral.
   • Human responses can be more varied in tone, style, and emotion, including informal language, colloquialisms, and typos.

2. Repetitiveness:

   • AI may repeat similar phrases or ideas within the same response, often because it's pulling from a large pool of information and trying to be thorough.
   • Humans are less likely to repeat themselves unless emphasizing a point for clarity.

3. Complexity of Thought:

   • AI can give comprehensive and well-organized answers, but sometimes it may lack deeper critical thinking, nuance, or personal experience.
   • Humans may express more subjective, personal, or complex opinions, backed by personal experiences or unique insights.

4. Contextual Awareness:

   • AI might struggle with nuances, local slang, and deep contextual awareness of the situation. For example, it might misinterpret emotions or offer advice that seems generic.
   • Humans are better at interpreting these subtleties and may provide more context-specific or empathetic responses.

5. Answering the Same Way Every Time:

   • AI may generate similar answers when asked the same or similar questions. It might provide a generic response that doesn't evolve much over time unless specifically retrained.
   • Humans often bring in their personal history, emotions, and unique perspectives, leading to slight variations in answers.

In terms of the original post, it's often easier to spot AI responses in writing that seems overly neutral or polished. If the response lacks a personal touch or specific anecdotal examples, it might be AI-generated. If you're asking about a particular example, feel free to share the text, and I can help analyze it further!

And that is from the horse's mouth (after feeding the quoted post into ChatGPT)

I must admit the poster concerned does have a very strange way of posting and I have often though their replies were AI generated.

GingerTim

QrizB said:

Brie said:

booneruk said:

Brie said:

But for AI replies - how can you spot that's what they are?  As opposed to someone just typing a bit about what they think they know??  I don't know what is from AI in the example shown in the original post.

There are plenty of AI text detectors that usually do a great job: https://www.google.com/search?q=ai+text+detector&oq=ai+text+detector

...but most of the time you can tell

can you give me an example of what the "tell" is on the example in the initial post?

Or maybe we can just ask @FlaatusGoat whether this reply or theirs was Ai-assisted or not?

https://forums.moneysavingexpert.com/discussion/comment/81180465/#Comment_81180465

I've read and dealt with a number of AI-generated essays in the last few months and I'm absolutely certain this is an AI response: https://forums.moneysavingexpert.com/discussion/comment/81186645/#Comment_81186645

Most of the usual tells: not entirely useful generalities, 'reach out', odd tone and expression in which no human writes, and so on.

born_again

Reach out.
Is a well used phrase by US customer support.

Please Reach Out to us with your problem.

GingerTim

born_again said:

Reach out.
Is a well used phrase by US customer support.

Please Reach Out to us with your problem.

Yes - but not particularly common in the UK, and ChatGPT etc are very US-centric.