If I had to guess it will go poorly. Having tried to use ChatGPT for parts of my job, it gets a lot subtly wrong. If I had just let it do my job, it would make a mess.
My job falls under IS, and in our monthly townhall we had a brief from the head of technology on our companies AI policy.
Our policy states we can only use approved AI systems, not sure which one but we have an account with one of the big ones. The main thing was, absolutely 100% of everything generated has to be reviewed and approved by a person before being used.
Basically any person using it will be on the hook as if they had created it themselves. They encouraged us to use it and explore where it can help, but made sure we knew we’d be responsible for the results.
Personally, I’d love to play around with it, as I do both graphics and programming as part of my job. The only issue I have is the language is so niche I doubt it’ll have enough data to generate anything, and I don’t want to feed proprietary code into an unsecured database to train it.
I’m sure in the next two years I’ll be using it for something.
Support Staff costs money. When they do their job right, they cost more money (refunds, repairs, etc…) . They don’t make money. The only reason support even exists is to keep the PR around the company good. If a company is small enough (or it is soon to be non-existent), it could just get rid of support and just hire a good PR firm and no one would know the difference.
Pretty easy to see that this will lead to immediate short-term gains but long-term pain. I can’t tell you the amount of infinite loops I’ve found myself in with AI chat, even among the most simplistic questions.
CEO, this year: Balance sheets look great! labor costs way down!
CEO, years out: Looks like profits are going down as people realize our products are shit., Better fire more people. Maybe I’ll sell off some assets too!
The CEO role is mostly about doing what you are told by shareholders. Last thing you want to do is remove the final limitations of a conscience from the position. An AI for ceo’ing will be optimised to be evil. More evil than the average psychopathic/narcissistic/Machiavellian CEO.
edit: Downvoters might wanna learn to read more thoroughly. I’m saying CEOs have almost no conscience. I’m not saying they have more than a narc/psycho/machi level of conscience, I’m saying they have a tiny tiny scrap of it(rarely), which the AI will not have.
… there’s no way their tech is good enough to avoid the billion and one ways to manipulate and confuse current-gen LLMs…
Unless he went with older tech, just feeding specific pre-programmed responses to specific keywords. That’s a glorified phone tree though, nothing really new.
There’s an in between. Have the LLM map into predefined responses, thus leveraging the LLM’s insane parsing abilities with the linearity of tree-like structures.
and doesn’t receive any wages or sick leave, obviously.
Well, not exactly, but these AI response systems are not cheap to run, you either:
A: Need to use a third party which charges you a non-trivial amount of money per response
B: Host it yourself on the cloud, which charges you a non-trivial amount of money for computation costs
C: Host it yourself bare metal, which costs a VERY non-trivial amount of front for hardware and still costs you a nontrivial (but lower amount) monthly for electricity/maint
But yeah no… they are a shit tonne cheaper than people though. So it is that the horse is being replaced by the streetcar.
I’ll be honest, if AI works as promised, I’d rather deal with AI than most support staff.
I mean how many times have people flat out told me they’d do something and didn’t, lied to me, didn’t know their policy (or claimed they didn’t know) ignored the language on the website, or refused to do something, which is over turned once you get a manager or a different rep.
Add long wait times on phones, and support staff that seem to disappear on calls… I mean damn.
Yes part of this is training, support, and policies, but a big part of this is support is made to be intentionally unhelpful because they know the more support people have to go through the more likely they will just accept it and move on.
I don’t see how AI is going to solve any of that if the intended result is to make getting actual support as frustrating as possible. If anything adding AI to the mix will just make things even more frustrating–but simultaneously also harder to pin down why so it could be stopped.
AI won’t get the resolution of stuff wrong or not know how to do X. AI will be able to do almost every task faster and shouldn’t leave people on hold while they go check stuff or such.
AI will be consistent, and thus we’ll see exactly how these companies want to handle it.
My assumption is AI is a tier instead of a full replacement, instead of 100 customer service, have 10 who handle the hard problems AI didn’t solve, and AI triages/fixes the generic problems.
Half the time I call my cable company they reboot my modem and it fixes something… That should be just an automatic option. Hell the “I can’t see your modem, we are opening a ticket to find out if there’s an outage in your are” again could be done by AI, and likely faster.
Really I see AI as a tool, and yeah, companies aren’t going to use it well, and will continue to be shitty, but ultimately the problem is we need to stop doing business with shitty companies.
But following that up, we need more options, I have ONE fucking cable provider in my home, so I’m stuck with (Bag of )Cox
I love people cosplaying as CEOs who clearly love being called a CEO. In reality, they have a product that’s likely a clone of an existing better product and barely have any employees.
You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: !technology@lemmy.world
This is a most excellent place for technology news and articles.
If I had to guess it will go poorly. Having tried to use ChatGPT for parts of my job, it gets a lot subtly wrong. If I had just let it do my job, it would make a mess.
My job falls under IS, and in our monthly townhall we had a brief from the head of technology on our companies AI policy.
Our policy states we can only use approved AI systems, not sure which one but we have an account with one of the big ones. The main thing was, absolutely 100% of everything generated has to be reviewed and approved by a person before being used.
Basically any person using it will be on the hook as if they had created it themselves. They encouraged us to use it and explore where it can help, but made sure we knew we’d be responsible for the results.
Personally, I’d love to play around with it, as I do both graphics and programming as part of my job. The only issue I have is the language is so niche I doubt it’ll have enough data to generate anything, and I don’t want to feed proprietary code into an unsecured database to train it.
I’m sure in the next two years I’ll be using it for something.
Support Staff costs money. When they do their job right, they cost more money (refunds, repairs, etc…) . They don’t make money. The only reason support even exists is to keep the PR around the company good. If a company is small enough (or it is soon to be non-existent), it could just get rid of support and just hire a good PR firm and no one would know the difference.
Pretty easy to see that this will lead to immediate short-term gains but long-term pain. I can’t tell you the amount of infinite loops I’ve found myself in with AI chat, even among the most simplistic questions.
Hell, I’ve caught myself in some of those infinite loops with actual people.
CEO, this year: Balance sheets look great! labor costs way down!
CEO, years out: Looks like profits are going down as people realize our products are shit., Better fire more people. Maybe I’ll sell off some assets too!
repeat until bankrupt
We’d probably be better off replacing 90% of CEOs with AI.
Would probably save more money.
Please no. :(
The CEO role is mostly about doing what you are told by shareholders. Last thing you want to do is remove the final limitations of a conscience from the position. An AI for ceo’ing will be optimised to be evil. More evil than the average psychopathic/narcissistic/Machiavellian CEO.
edit: Downvoters might wanna learn to read more thoroughly. I’m saying CEOs have almost no conscience. I’m not saying they have more than a narc/psycho/machi level of conscience, I’m saying they have a tiny tiny scrap of it(rarely), which the AI will not have.
… there’s no way their tech is good enough to avoid the billion and one ways to manipulate and confuse current-gen LLMs…
Unless he went with older tech, just feeding specific pre-programmed responses to specific keywords. That’s a glorified phone tree though, nothing really new.
There’s an in between. Have the LLM map into predefined responses, thus leveraging the LLM’s insane parsing abilities with the linearity of tree-like structures.
It can be done, and can be done intelligently.
Which I somewhat doubt was the case here.
This was always going to happen. As is the revolution that will eventually overthrow them. Capitalism isn’t built for sustainability.
I hate AI as it is now. I really do.
Why do you hate AI instead of hating business schools culture? AI will save lives, ceos use it to fire people
You’re right. Apologies.
Wow an admission of being incorrect on the internet. I must be dreaming
Well,you’re not x) Hope everything is okay with you. Cheers x
Well, not exactly, but these AI response systems are not cheap to run, you either:
A: Need to use a third party which charges you a non-trivial amount of money per response
B: Host it yourself on the cloud, which charges you a non-trivial amount of money for computation costs
C: Host it yourself bare metal, which costs a VERY non-trivial amount of front for hardware and still costs you a nontrivial (but lower amount) monthly for electricity/maint
But yeah no… they are a shit tonne cheaper than people though. So it is that the horse is being replaced by the streetcar.
I’ll be honest, if AI works as promised, I’d rather deal with AI than most support staff.
I mean how many times have people flat out told me they’d do something and didn’t, lied to me, didn’t know their policy (or claimed they didn’t know) ignored the language on the website, or refused to do something, which is over turned once you get a manager or a different rep.
Add long wait times on phones, and support staff that seem to disappear on calls… I mean damn.
Yes part of this is training, support, and policies, but a big part of this is support is made to be intentionally unhelpful because they know the more support people have to go through the more likely they will just accept it and move on.
I don’t see how AI is going to solve any of that if the intended result is to make getting actual support as frustrating as possible. If anything adding AI to the mix will just make things even more frustrating–but simultaneously also harder to pin down why so it could be stopped.
AI won’t get the resolution of stuff wrong or not know how to do X. AI will be able to do almost every task faster and shouldn’t leave people on hold while they go check stuff or such.
AI will be consistent, and thus we’ll see exactly how these companies want to handle it.
My assumption is AI is a tier instead of a full replacement, instead of 100 customer service, have 10 who handle the hard problems AI didn’t solve, and AI triages/fixes the generic problems.
Half the time I call my cable company they reboot my modem and it fixes something… That should be just an automatic option. Hell the “I can’t see your modem, we are opening a ticket to find out if there’s an outage in your are” again could be done by AI, and likely faster.
Really I see AI as a tool, and yeah, companies aren’t going to use it well, and will continue to be shitty, but ultimately the problem is we need to stop doing business with shitty companies.
But following that up, we need more options, I have ONE fucking cable provider in my home, so I’m stuck with (Bag of )Cox
If they can solve the honesty problem with AI maybe.
I love people cosplaying as CEOs who clearly love being called a CEO. In reality, they have a product that’s likely a clone of an existing better product and barely have any employees.