AI will be the end of forums like AFF, FlyerTalk etc

[markis10]

Maybe not for a while….

 

[Flyerwin]

I doubt it. I come here to lurk the forums and watch people furiously agree and prove that they are right to one another. AI can’t provide that form of passive entertainment.

Plus I always learn something random and new each time I visit. Can’t beat that.

 

[Gold Member]

Theoretically, AI doesn’t have personal and political biases that would unfairly punish valid commentary. It would surely be a shame if that day came to pass here on AFF.

 

[marki]

Theoretically, AI doesn’t have personal and political biases that would unfairly punish valid commentary. It would surely be a shame if that day came to pass here on AFF.

Theoretically correct and agree. But AI does draw on human info. So is therefore biased. We would have Ai talking to AI? very engaging ( Not )

 

[Gold Member]

Theoretically correct and agree. But AI does draw on human info. So is therefore biased. We would have Ai talking to AI? very engaging ( Not )

That would be a worthwhile enhancement.

 

[marki]

That would be a worthwhile enhancement.

How would one realise that AI is chatting to AI?

 

[Gold Member]

How would one realise that AI is chatting to AI?

The conveniently worded AFF Terms of Service conveniently prevent me from speaking my mind, but I think you get my gist.

 

[marki]

The conveniently worded AFF Terms of Service conveniently prevent me from speaking my mind, but I think you get my gist.

There is no explicit mention of AI in the context of the AFF's terms of service when I did the search. However, given the modern context where AI might be used for moderation, content suggestion, or other backend processes, one can infer:
AI for Moderation: The mention of staff approval for comments suggests a human element in moderation, but doesn't rule out AI assisting in filtering content for spam or inappropriate material!

 

[odysseus]

You can see it happening already on X (twitter), where many 'train' their AI bots. Some article/story will be posted. Then another poster will post basically the same content, but reworded in AI speak.

Sometimes you do see them try to engage in a discussion. As useless as standard AI, but it takes up space. And then there are the times where they break, and you just get garbage posted, or people cause them to post garbage by 'breaking' their rules.

It's certainly much more dry, and repetitive compared to genuine content.

Also, in respect of the poster above talking about theoretically not having bias, that's when prepared by a relatively neutral entity. Just like traditional media where there was more objectivity. But once deployed by private individuals they are often set with an agenda, which may include a strong political purpose.

 

[marki]

You can see it happening already on X (twitter), where many 'train' their AI bots. Some article/story will be posted. Then another poster will post basically the same content, but reworded in AI speak.

Sometimes you do see them try to engage in a discussion. As useless as standard AI, but it takes up space. And then there are the times where they break, and you just get garbage posted, or people cause them to post garbage by 'breaking' their rules.

It's certainly much more dry, and repetitive compared to genuine content.

Also, in respect of the poster above talking about theoretically not having bias, that's when prepared by a relatively neutral entity. Just like traditional media where there was more objectivity. But once deployed by private individuals they are often set with an agenda, which may include a strong political purpose.

 

[smiliemonster]

Theoretically correct and agree. But AI does draw on human info. So is therefore biased. We would have Ai talking to AI? very engaging ( Not )

And the biases of the people who code the AI tool/platform to start with. Add the information and its inherent biases on top, with a sprinkling of assertive arrogance (i.e. the results are presented as though they are fact), and many won't realise the output is wrong until they're affected.

Mind you, people in forums like this can offer the wrong information from time to time too, but here there's an in-built peer review process (so to speak) that doesn't exist in AI tools at the moment. If someone tells you the wrong thing on AFF, others with more knowledge and experience will help set things straight. To me, that's a fabulous and valuable resource.

 

[MEL_Traveller]

View attachment 427074

Maybe not for a while….

not just qantas/EK… you search ‘is sichuan airlines a member of star alliance’ and the answer comes back a resounding ‘yes’.

in fact, type in just about any Chinese airline and ask if it is star alliance and it will tell you the same thing!

I’m sure AI Will improve over time but at the moment in the travel sector it is a bit rubbish!

 

[Forg]

not just qantas/EK… you search ‘is sichuan airlines a member of star alliance’ and the answer comes back a resounding ‘yes’.

in fact, type in just about any Chinese airline and ask if it is star alliance and it will tell you the same thing!

It makes me wonder, albeit in a "not really interested enough to spend time on it" kinda way, the source of the incorrect info. It can go to the horses' mouth and get the correct info from numerous airline sites as well as the OneWorld site, and it probably does/has, but in order to come back with the wrong answer there must be a bunch of info it's found which is incorrect.

Maybe - 'cos the other possibility is it just doesn't understand the actual question & is behaving like an annoyingly large number of new-hires do & just pretending it understands because it thinks it's better to appear like it understands & give the wrong answer than to appear it doesn't understand by asking for clarification ...

 

[MEL_Traveller]

It makes me wonder, albeit in a "not really interested enough to spend time on it" kinda way, the source of the incorrect info. It can go to the horses' mouth and get the correct info from numerous airline sites as well as the OneWorld site, and it probably does/has, but in order to come back with the wrong answer there must be a bunch of info it's found which is incorrect.

Maybe - 'cos the other possibility is it just doesn't understand the actual question & is behaving like an annoyingly large number of new-hires do & just pretending it understands because it thinks it's better to appear like it understands & give the wrong answer than to appear it doesn't understand by asking for clarification ...

I did a bit of digging… Sichuan Airlines for example was mentioned deep on some star alliance page as having code share agreements for some connecting flights! So the AI picked up on that and assumed it was part of the alliance.

I guess there’s not a lot of interest out there to correct the AI for airline alliances atm! But interestingly I just googled China Eastern premium economy a few minutes ago… the top AI result shows CE’s first class seats.

Looking for a day’s outing (a self tour itinerary) for an upcoming holiday, AI recommended eating at a couple of restaurants that don’t exist!

Still a ways to go!

 

[moa999]

The AI models are always super confident.

But you've got to wonder what exactly causes that issue given I'd think their are thousands of sites that name all the one world members.

 

[TonyHancock]

Pffft and I am just in the process of becoming a travel Influencer with my "Tales of a Misanthrope" blog.

My "TukTuk" content is ready to go.

 

[Forg]

If I hadn’t met you & didn’t know you so well, @TonyHancock, I’d have suspected that wasn’t the photo I took last week & was generated by AI!!

I know how touchy you are about your hands so I was surprised you even let me take it.

 

[Hvr]

Pffft and I am just in the process of becoming a travel Influencer with my "Tales of a Misanthrope" blog.

My "TukTuk" content is ready to go.

I can see you loving AI until they expose your love for France and all things French .