Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Chatbots: Still dumb after all these years

mindmatters.ai

Post reply on HN
Print view View original

21–30 of 426 posts

Re: Chatbots: Still dumb after all these years

#21

Post by reaperducer »

I've never had a good experience with a chatbot. They're the equivalent of waiting on hold on the telephone with your bank and the recording asking you over and over if you know that you can accomplish some simple and completely unrelated task on the web site.

It's even worse when chatbots pretend to be real people, though I've seen less of that lately.

The Apple Card chatbot is one of the worst I've used. This is an actual conversation when I tried to dispute an unknown trasnsaction (I screenshotted it for posterity):

Apple: I understand how frustrating this could be for you. I'm going to connect you with a specialist that will review your account with you and provide more specific details to resolve this inconvenience.

Apple: Hello, I will be glad to assist you with that today. Allow me a few moments to review your account.

Me: It's been 15 minutes. Are you still there?

Apple: Welcome back. Let me know how I can assist you further.

Me: What do you mean welcome back? I've been waiting for you.

Apple: Sure thing. I will be here when you get back.

Re: Chatbots: Still dumb after all these years

#22

Post by reaperducer »

benbristow wrote: post #9
arthur_sav wrote: post #5

I really hate when companies hide behind chatbots. Like, NOOOO, your "AI" chatbot didn't help me - get me an actual person.

It can be quite useful in a weird way. While everyone else is stuck on the chatbot going round in circles, usually typing something like "human" or "talk to human" will make the chatbot connect you with an actual human representative (or ask you a few basic questions first and then connect you). I've used the trick on various large company's websites when trying to get support and it seems to be quite 'universal'.

Reminds me of how I used to occasionally see people on the bus shouting into their phones "OP-ER-AY-TOR!" "REP-RE-SENT-UH-TIVE!" "HUE-MAN!"

Re: Chatbots: Still dumb after all these years

#23

Post by hotpotamus »

klyrs wrote: post #12

My favourite flaw of chatbots is exposed by ELIZA. Not chatting with ELIZA, (though, it does suffer this flaw) but using responses inspired by that program. "Please elaborate on that" or "tell me more about [noun]" etc. Bots appear to have zero lines of short term memory, and utterly fail to pick up a reference to the thing that they just said. My favorite being bot> [something plausibly human-sounding] me> What do y…

I'm not trying to be mean or break any HN rules, but did you read the article? It basically covers what you asked and was quite a revelation to me. Others in here point out that GPT-3 is not a chatbot which is good info, but I also wonder if there is anything out there that can even "remember" the context of a conversation it's supposed to be having.

Particularly interesting is the question from the article, "who is the current president". Seems like a slam dunk but it could not be answered. Interestingly this is a common question paramedics will give to people who have suffered head injuries to assess consciousness.

Re: Chatbots: Still dumb after all these years

#24

Post by reaperducer »

codegeek wrote: post #17
arthur_sav wrote: post #5

I really hate when companies hide behind chatbots. Like, NOOOO, your "AI" chatbot didn't help me - get me an actual person.

We have a chatbot on our marketing site but it says something like "I am a Bot. Once I ask you a couple of questions, I can connect you to a human. Is that ok ?".

What happens if someone responds with, "No?"

Re: Chatbots: Still dumb after all these years

#25

Post by chasd00 »

codegeek wrote: post #17
arthur_sav wrote: post #5

I really hate when companies hide behind chatbots. Like, NOOOO, your "AI" chatbot didn't help me - get me an actual person.

We have a chatbot on our marketing site but it says something like "I am a Bot. Once I ask you a couple of questions, I can connect you to a human. Is that ok ?".

i like that better than fighting with a bot to trigger an actual human getting online. However, a few short questions and connect to human sounds like a form and a submit button.

Re: Chatbots: Still dumb after all these years

#26

Post by dmkirwan »

While the tech is constantly advancing, I have yet to use one in anger that fully solved my problem outside of giving me information that's already easy to find.

I think the only place it makes sense is in very very specific use-cases. Think about, say, a dental clinic. I would guess that if you put a human on chat support for a year in that setting, you'd see common themes emerging in customer problems. Maybe you develop a good library of canned responses and you find they work for 70 - 80% of queries, great. Then you could probably build a good chat bot to handle questions and handle cancelling/rescheduling appointments.

Re: Chatbots: Still dumb after all these years

#27

Post by waynesonfire »

the_only_law wrote: post #16
arthur_sav wrote: post #5

I really hate when companies hide behind chatbots. Like, NOOOO, your "AI" chatbot didn't help me - get me an actual person.

Still doesn’t piss me off as much as automated phone support systems. 90% of the time they only offer options that I could easily go online to do. If I’m calling your phone number. It’s because I have a problem that’s not solved or clarified by your existing self-support systems.

> Still doesn’t piss me off as much as automated phone support systems.

Yes, the AI voice bot is marginally better because I can request "customer service" without waiting to discover the right numeric code. That's about the extent of that.

Re: Chatbots: Still dumb after all these years

#29

Post by pydry »

kapp_in_life wrote: post #18
the_only_law wrote: post #16

Earlier quoted context omitted.

Still doesn’t piss me off as much as automated phone support systems. 90% of the time they only offer options that I could easily go online to do. If I’m calling your phone number. It’s because I have a problem that’s not solved or clarified by your existing self-support systems.

Unfortunately I think they've found that most people don't(won't?) do that self debugging. Relevant xkcd: https://xkcd.com/806/

They could do a better job of separating the two.

Most companies would save their time and mine if they had a callback system which took as many details as possible up front & didnt have to ask my name and account details because Im already logged in.

Even better if it could give me some warning about where I am in the queue and when the call might be coming (e.g. via push notification).

I'd 1000x prefer that over any oversold AI hooked up to an FAQ.

Unfortunately it seems call centres are driven by very traditional metrics which wouldnt lead anybody to set up a system like this.

Re: Chatbots: Still dumb after all these years

#30

Post by klyrs »

hotpotamus wrote: post #23
klyrs wrote: post #12

My favourite flaw of chatbots is exposed by ELIZA. Not chatting with ELIZA, (though, it does suffer this flaw) but using responses inspired by that program. "Please elaborate on that" or "tell me more about [noun]" etc. Bots appear to have zero lines of short term memory, and utterly fail to pick up a reference to the thing that they just said. My favorite being bot> [something plausibly human-sounding] me> What do y…

I'm not trying to be mean or break any HN rules, but did you read the article? It basically covers what you asked and was quite a revelation to me. Others in here point out that GPT-3 is not a chatbot which is good info, but I also wonder if there is anything out there that can even "remember" the context of a conversation it's supposed to be having. Particularly interesting is the question from the article, "who is…

Hidden in my comment is a question: has anybody even tried to included previous lines into the prompt, and does that not help? Asking here because there are a lot of ML nerds who could probably do a proof of concept in a few hours, if not a few lines of code, so there's a decent chance I'll get an informed response.
Post reply on HN