You may be thinking that the CS response was a little odd - this might explain why.

Spoonie_Turtle

CKhalvashi said:

Spoonie_Turtle said:

CKhalvashi said:

CSI_Yorkshire said:

Is it just me, or do some of the new posters on the forum also sound a bit AI?  Several lengthy posts in rapid succession, across a wide range of boards, all sounding a little 'clunky' and using unusual but identically repeated phrasing.

Noticed this also, especially those at early hours.

I'm not sure also though if those who are new to the forum are using some form of AI now it's freely available to assist in writing posts for help to make sure they don't leave anything out, so all may not be as it seems.

If they have to use a script bot to post, what's the point?  That's not at all the same as looking things up to check and providing sources for advice they want to give themselves.  If they're using a script bot to give advice they don't fully know about or understand, then they'd be better off just telling the original poster to Google it rather than giving potentially misleading advice.

… at the opposite end of the scale many just don't possess the skills to be able to adequately detail a problem they aren't familiar with and don't deal with terminology-wise on a daily basis.

I understand your point but on a forum like this, people should be giving advice that they know and understand.  If they don't know enough to be able to look up the term (in English, in this case) or to be able to manually use Google translate, they probably don't know enough to be giving advice.

One particular account in question appearing in the linked thread above sounds VERY much like an AI has been asked to write advice for X topic.  It's not a case of sounding odd because English is not their first language, or because they're code-switching, it's actually too slick and impersonal, as if addressing a wide audience in an article rather than addressing individual posters and the specific situation at hand.

markin said:

I would barely be ably to post without a spellcheck, That is machine learning is it not? So wouldn't blame anyone for using a tool to help them articulate what they want to say better.

Some disabled people will be using speech to text, i would expect that to look odd at times, Often the poster will warn they are using it.

There's using a tool to help, and then there's using it to write a whole post that doesn't even necessarily actually help.  And see my observations above as to why these recent ones don't sound right.

Speech to text is totally different from chat bots and not at all in question here.

markin said:

It says all responses are proof read so i would expect the names to be real, For now anyway, As they don't have to find a standard response to paste or think about and write a reply its going to be 90% faster.

Was this about Octopus?  If they proof-read the response then they definitely didn't read the original query to assess whether the response was appropriate.

Spoonie_Turtle

Update on my 'conversation' with Octopus: AI says no, to a Direct Debit level a human approved last winter before the automatic £67 increase.

Friday late afternoon was clearly not a good time to ask, not sure whether to bother trying again in the week during office hours instead.

QrizB

CSI_Yorkshire said:

Spoonie_Turtle said:

CKhalvashi said:

CSI_Yorkshire said:

Is it just me, or do some of the new posters on the forum also sound a bit AI?  Several lengthy posts in rapid succession, across a wide range of boards, all sounding a little 'clunky' and using unusual but identically repeated phrasing.

Noticed this also, especially those at early hours.

I'm not sure also though if those who are new to the forum are using some form of AI now it's freely available to assist in writing posts for help to make sure they don't leave anything out, so all may not be as it seems.

If they have to use a script bot to post, what's the point?  That's not at all the same as looking things up to check and providing sources for advice they want to give themselves.  If they're using a script bot to give advice they don't fully know about or understand, then they'd be better off just telling the original poster to Google it rather than giving potentially misleading advice.

https://forums.moneysavingexpert.com/discussion/6450633/moved-to-new-property-any-particular-company-for-gas-and-electricity#latest

The posting history of a brand new member in this thread is as you say.  I can't see generated advice being up to date, useful or specific enough to issues that develop in discussion.

I'm wondering if it's spammers.

Usually, spammers get short shrift here since new account can't post links until they've made a few posts.

By using LLM AI to generate plausible-sounding answers, a spammer could keep the account alive (without needing to put in much actual work or know anything about the subject) for long enough to unlock link-posting rights.

Of course, this could just me being cynical after previous spells as a forum moderator! (Not this forum, but elsewhere.)

markin

Spoonie_Turtle said:

CKhalvashi said:

Spoonie_Turtle said:

CKhalvashi said:

CSI_Yorkshire said:

Is it just me, or do some of the new posters on the forum also sound a bit AI?  Several lengthy posts in rapid succession, across a wide range of boards, all sounding a little 'clunky' and using unusual but identically repeated phrasing.

Noticed this also, especially those at early hours.

I'm not sure also though if those who are new to the forum are using some form of AI now it's freely available to assist in writing posts for help to make sure they don't leave anything out, so all may not be as it seems.

If they have to use a script bot to post, what's the point?  That's not at all the same as looking things up to check and providing sources for advice they want to give themselves.  If they're using a script bot to give advice they don't fully know about or understand, then they'd be better off just telling the original poster to Google it rather than giving potentially misleading advice.

… at the opposite end of the scale many just don't possess the skills to be able to adequately detail a problem they aren't familiar with and don't deal with terminology-wise on a daily basis.

I understand your point but on a forum like this, people should be giving advice that they know and understand.  If they don't know enough to be able to look up the term (in English, in this case) or to be able to manually use Google translate, they probably don't know enough to be giving advice.

One particular account in question appearing in the linked thread above sounds VERY much like an AI has been asked to write advice for X topic.  It's not a case of sounding odd because English is not their first language, or because they're code-switching, it's actually too slick and impersonal, as if addressing a wide audience in an article rather than addressing individual posters and the specific situation at hand.

markin said:

I would barely be ably to post without a spellcheck, That is machine learning is it not? So wouldn't blame anyone for using a tool to help them articulate what they want to say better.

Some disabled people will be using speech to text, i would expect that to look odd at times, Often the poster will warn they are using it.

There's using a tool to help, and then there's using it to write a whole post that doesn't even necessarily actually help.  And see my observations above as to why these recent ones don't sound right.

Speech to text is totally different from chat bots and not at all in question here.

markin said:

It says all responses are proof read so i would expect the names to be real, For now anyway, As they don't have to find a standard response to paste or think about and write a reply its going to be 90% faster.

Was this about Octopus?  If they proof-read the response then they definitely didn't read the original query to assess whether the response was appropriate.

It was posted back on the 10th and looked like some of you had forgotten.


"A representative for Octopus Energy told Insider: "Our team supervises the answers AI provides, so, for example, drafting a personalized response that a team member can review and then send on."

Workers at other companies recently reported on the experience of using AI tools to help carry out their work. One worker told Vice that OpenAI's ChatGPT did 80% of their job.

Jackson told the Times of London that the development was unlikely to lead to job losses at his company but that the pace of AI technology had the potential to cause "huge and rapid dislocation" to the job market."

the_lunatic_is_in_my_head

CSI_Yorkshire said:

Is it just me, or do some of the new posters on the forum also sound a bit AI?  Several lengthy posts in rapid succession, across a wide range of boards, all sounding a little 'clunky' and using unusual but identically repeated phrasing.

How do we know you aren't AI creating a diversion to to draw the spotlight elsewhere  

markin

the_lunatic_is_in_my_head said:

CSI_Yorkshire said:

Is it just me, or do some of the new posters on the forum also sound a bit AI?  Several lengthy posts in rapid succession, across a wide range of boards, all sounding a little 'clunky' and using unusual but identically repeated phrasing.

How do we know you aren't AI creating a diversion to to draw the spotlight elsewhere  

80% of internet traffic is bots!

CSI_Yorkshire

the_lunatic_is_in_my_head said:

CSI_Yorkshire said:

Is it just me, or do some of the new posters on the forum also sound a bit AI?  Several lengthy posts in rapid succession, across a wide range of boards, all sounding a little 'clunky' and using unusual but identically repeated phrasing.

How do we know you aren't AI creating a diversion to to draw the spotlight elsewhere  

I'd have to be an advanced model, and nobody has ever accused me of being advanced!