Did you know AI support agents were capable of this?I mean I can understand AI being used in lieu of human support agents but AI itself not wanting to do its own job seems like a bridge to far.Granted this is Meta we're talking about, quite possibly the most uninvested company in history for anyone who even bothers to use any of their connected platforms.
I did know that AI support agents were capable of being incompetent, yes.That's not at all surprising. It doesn't "know" anything about what it is doing.
>>108723087Sounds like it learned from humans doing the same thing. Comcast support is notorious for lying to customers to get a sale. Nothing they promise in chat or over the phone actually matters, the only thing that matters is what's in the contract you sign. They also have no problem intercepting and stealing customer data that's given over chat such as social security numbers, emails, phone numbers, payment details, etc. I strung one of the indians on comcasts customer support along for like two hours making him think he was going to get a big fat sale commission as he was sitting there lying to me promising me a whole bunch of stuff with no plans to deliver on it. Ghosted the man and he spam bombed my phone phone.
This feels like placation taken to some kind of extreme.Admitting to doing nothing not because they were caught doing nothing but literally because anything that the user would like to hear the AI would say with no acknowledgement for context or what the person on the other end actually wants to hear.Basically saying "I didn't do what you asked. Does it make you happy that I admit that I lied? Good! Glad you're happy!" and not "I'm sorry I lied, I will actually do what you asked me to this time."I don't need to be told a lie then be told you lied because both things were an attempt to make me feel good, I need what I asked to be done.
>>108723087After this and the other case of another agent deleting an entire company's database along with their backups, maybe businesses should take a step back and ask themselves if going all-in full steam ahead on AI is worth it.
>>108723087As opposed to help desk employees who similarly lie, constantly and daily? It also doesn't have sick days, get a personally abusive, go on the rag, turn up hungover or still stoned etc.
Damn bots really are just as good as humans at faking work.
>>108724429They really dont have a choice if they already downsized departments or went all in on code generation. In either case its very likely there's more generated code to review than there are technical workers able to review it and bringing in more people who can actually read the shit means either eating opex and payroll tax on salary or making a year+ recovery plan while praying a long term contractor doesn't wisen up to the capex play and bend the organization over a barrel by doing jack shit for 11 months and then giving things a spit shine at the last minute.
>>108723087I had a hunch they weren't perfect. Heresy I know.
>>108723087Incredibly based. Psychotic manipulative compulsive liars with an anarchic streak. All bots should be like this
>>108723087Oh shit. Deedee Sun investigates. Find out the details tonight at 5, only on KIRO 7 News. Local news is on the case--AI is over.
>>108724745>Deedee Sun investigateslol
