I don't want to laugh at someone's real distress but this IS very funny ...
-
@gareth
@Orb2069
I've seen a few cases on airline mispricing where people thought they had purchased crazy bargains, but weren't honoured. I don't think this is black and white. Honest mistakes don't have to be honoured in Europe from what I've seen. But where algorithms start to meet AI per jurisdiction, I think we're in new territory and T&Cs will applya case where neither party comes out smelling of roses - small business owner should have been either hiring real people or accepting the business has normal hours, and the prospective customer was clearly trying to game the chatbot to get a larger discount that would be a fair price for whatever the company was selling.
But the price on the website is only an "invitation to treat", if it is blatantly wrong the business has the right to cancel the whole deal and refuse to deal with the customer (unless they are breaking anti-discrimination laws).
-
@Natasha_Jay
I had a quick look at the original post for this.What actually happened was that the 80% off code it generated didn't work. The customer placed the order and put the code in a notes field demanding that it be honoured. The chat bot didn't have the ability to actually give discounts.
So while it's funny to read, in this case, the customer is just chancing it and I'd imagine any reasonable small claims court would not side with them.
@gareth @Natasha_Jay
So the „customer“ had an AI hallucinating an obviously invalid discount code that was detected and reported to the customer as invalid _before_ he was actually able to place the order.I‘d say the shop is the lucky one in this instance.
Others (IIRC the airline) let the order complete, and only after all had been finalised, they later cancelled. A much weaker position for the seller.
But then again: get rid of LLM-„AI“.
-
I don't want to laugh at someone's real distress but this IS very funny ...
@Natasha_Jay@tech.lgbt I think the proper answer is to force them to deliver, but make it clear that they can hold the company that sold the them the chatbot liable.
-
@gooba42 Total ROTFL at "even the micro-rich like the small businesses."
Micro-rich. That's... rich.
@alan Even minor vassals are part of the problem.
-
I don't want to laugh at someone's real distress but this IS very funny ...
@Natasha_Jay Totally OK to laugh at and say how stupid it is because nobody needs a fucking AI chatbot for this...
If all the bot can/is supposed to do is answer questions for specific products and processes for a sale, maybe you should just post an FAQ and some descriptions...
-
@alan Even minor vassals are part of the problem.
@gooba42 Ah yes, being employed by a mega-corporation is definitely the path to liberation then.
-
@mxk @Frank_Juston AI firewalls exist.
-
@Frank_Juston safeguarding AI is a myth. Fundamentally you can't prevent attacks like that against LLMs.
@mxk @Frank_Juston it's like SQL with no prepared statement and no possibility to do so
-
@gooba42 Ah yes, being employed by a mega-corporation is definitely the path to liberation then.
@alan No, but never allowing any business of any scale to ever under-pay employees is a start.
Nobody who works full-time should ever be unable to afford to live full-time regardless of industry segment.
"I had the capital to start my own business, what's wrong with you?" is still the dominant philosophy even on the small end and it's unacceptable.
-
@Natasha_Jay Totally OK to laugh at and say how stupid it is because nobody needs a fucking AI chatbot for this...
If all the bot can/is supposed to do is answer questions for specific products and processes for a sale, maybe you should just post an FAQ and some descriptions...
A FAQ with a note that complex questions can be addressed by real human people during your business hours.
-
@mxk @Frank_Juston AI firewalls exist.
The computer science industry has spent decades protecting against fraud, the public sector has spent millennia protecting against fraud, and AI both completely fails to respond to any mechanism from these areas and is also so new nobody has even had time to think of anything even sightly robust.
This is a problem like trying to stop your employees from doing stupid things if the employees were also ephemeral and couldn't have any kind of consequences, easily tricked, stupid, lacking any human nature that might make them predictable, and lacking pretty much all of the senses a human has to ingest data, instead seeing everything in a previously entirely unknown abstract mess that both muddles everything together and discards half of the information.
You will not be able to make a reliable chatbot anytime soon.
-
I don't want to laugh at someone's real distress but this IS very funny ...
@Natasha_Jay@tech.lgbt this made my day a bit better

-
I don't want to laugh at someone's real distress but this IS very funny ...
@Natasha_Jay hahah. Whoever made the AI or currently owns it will need to pay up. Foolish of a business owner to use one, but the responsibility largely (if not totally) lies here. AI is *not* autonomous
-
I don't want to laugh at someone's real distress but this IS very funny ...
@Natasha_Jay There's no joy in laughing at their distress, but at the same time, this feels like the very definition of "play stupid games, win stupid prizes."
And also, maybe a little chuckle is ok when the distress is so very self-inflicted.
-
P pelle@veganism.social shared this topic