AI Versus "Soul" A Study in Forum Tolerance For the Mundane

[PhotoHobbyist128 154]

All customer service/help desks typically follow a proprietory product/service oriented flow chart. Prior to AI we hired entry level people to work the CS/help desk, gave them the chart of canned responses, a phone and turned them loose. Speech recognition software started showing up in the 50's. I had Dragon in the 90s for speech to text. It was only a matter of time before Voice Cloning started in the late 90s and expanded rapidly to combine input and output. Using AI on the CS/help desk is a no brainer for companies, because it doesn't require subjective thinking or intuition. It's working off a canned response list, when it reaches the end of its options it has to transfer you to a live person.

It wasn't going off a canned script. We had a conversation. I asked questions and she answered as if human. That is what is scary.

 

[AlanKlein94 10]

Only partially correct as you've taken parts out of context. Ai forms artificial networks that "mimic" the brain's neural biological structures. AI forms modular based interconnected systems that replicate the neural pathways based on a mathematical model. AI lacks the ability to feel pain or pleasure, subjective consciousness, or genuine life experiences. It mimics human-like responses by processing vast amounts of data to emulate a thought, it can sound convincing but it's not actual self awareness. While it has the ability to learn based on evolving data, it still lacks intuition, and independent thought. Therein lies another of the big problems with AI, who's supplying the data. AI's ability to present a logical sounding thought, could be affected by tainted or manipulated data, misleading people.

Since it votes with the majority viewpoint, it can't be innovative and could often be wrong. How would you know? Minority viewpoints are discarded or minimized. It can't think out of the box. I don;t know if this is happening, but if AI is building on previous conclusions made by other AI systems which used prvious bad human data learning, they could be multiplying the errors and wrong thinking taking us down a back alley. Allowing them to operate on their own without regular human review would be foolish.

 

[AlanKlein94 10]

Recently, I had deal with a utility company and a recording said I would transferred to an AI agent. It was indistinguishable from a real human having proper intonation, timing, conveyed emotion, and even slight pauses to convey a thought process, but no delays in responses. If I hadn't been told beforehand it was AI I would not have know it was not a real human being. It is getting scary.

Ask it if you should sign a end of life Do Not Resuscitate form. The point is, these systems are set up with a list of normal and usual questions and give normal and applicable answers that people ask a utility company. It's a lookup table. Sure. It sounds human. But that's just an advancement in data-to-human voice algorithms, not intelligence. Don't be fooled. The machine is stupid.

 

[smoke665]

Since it votes with the majority viewpoint, it can't be innovative and could often be wrong. How would you know? Minority viewpoints are discarded or minimized. It can't think out of the box. I don;t know if this is happening, but if AI is building on previous conclusions made by other AI systems which used prvious bad human data learning, they could be multiplying the errors and wrong thinking taking us down a back alley. Allowing them to operate on their own without regular human review would be foolish.

Interesting, I hadn't really considered this, but if my understanding of how the learning process occurs is correct then your assumption would be correct. Even scarier if AI were to discover politicians and media's penchant for stretching the truth!

 

[AlanKlein94 10]

Interesting, I hadn't really considered this, but if my understanding of how the learning process occurs is correct then your assumption would be correct. Even scarier if AI were to discover politicians and media's penchant for stretching the truth!

Exactly. Just like social media, conclusions and beliefs are built within the group that believes a certain way. The same beliefs are repeated and carried forward. Confirmation bias seems to be built into AI as well. GIGO. Ignorant teachers. make ignorant pupils. For AI to be truly successful, there must be a way to eliminate repeating mistakes of the past that it has learned. Even with expert data, it still will repeat just what's known. Light bulbs do not go off in a machine.

 

[smoke665]

Confirmation bias seems to be built into AI as

That has been a concern for sometime.

Also it seems that AI has a BS factor built in, it's designed to predict the next word or sequence based on observed patterns. However the goal is to generate plausible content, not to verify the truth, so accuracy in the output is often coincidental and might produce content that sounds reasonable but is inaccurate. In other words if it doesn't know it'll make something up. My son is an IT professional who's mentioned this several times.

 

[AlanKlein94 10]

That has been a concern for sometime.

Also it seems that AI has a BS factor built in, it's designed to predict the next word or sequence based on observed patterns. However the goal is to generate plausible content, not to verify the truth, so accuracy in the output is often coincidental and might produce content that sounds reasonable but is inaccurate. In other words if it doesn't know it'll make something up. My son is an IT professional who's mentioned this several times.

This is probably OK for most online general use, when people are looking for most probable answers to existing knowledge. One would hope, that AI used in engineering, design, and advanced scientific use, the programs are set up to meet more stringent requirements. Fill in the blanks and guessing won;t be allowed unless flagged as out of the box "thinking". . Minority viewpoints will be given some authority,

 

[smoke665]

This is probably OK for most online general use, when people are looking for most probable answers to existing knowledge. One would hope, that AI used in engineering, design, and advanced scientific use, the programs are set up to meet more stringent requirements. Fill in the blanks and guessing won;t be allowed unless flagged as out of the box "thinking". . Minority viewpoints will be given some authority,

Actually its a fairly well know problem. There have been some misguided attorneys who've used AI to write briefs, only to get in court and find it quoted nonexistent cases. Technical and research writers have found it has made up sources. A Typical AI model isn't assessing whether the information it provides is correct. Its goal when it receives a prompt is to generate what it thinks is the most likely string of words to answer that prompt. Sometimes this results in a correct answer, but sometimes it doesn’t – and the AI cannot interpret or distinguish between the two. Kind of scary, and negates the time savings if you have to proof read carefully what it generates.

 

[TJMcG125 8]

There has been some kind of Internet glitch that has blocked most of the web from me all day. As a result, I have been online with the bots (Google, Gemini, ChatGPT) all day trying to find out what's wrong here. I can access TPF, Facebook, YouTube, my Yahoo email, my Gmail, and others, but many sites and links don't work. I can get Adobe's website but not the Creative Cloud or Generative servers.

I spent the morning trying all the remedies they gave me, but all of this is very foreign to me, and the last thing I need is to be making changes to Windows or to my computer that I don't understand and have to take it in to get it fixed. Verizon had a major failure on the 14th-15th that is supposed to be resovled, but now the bots are telling me that there is some kind of data logjam in Phoenix and Tucson that is still causing my problems. They assure me thatI have done everything I can and that there is nothing more I can do but wait it out. I hope that's right, but I can't trust the AI yet.

Thing is, I don't know anything about this stuff myself, and I'm not sure that being reassured that it's all Verizons's fault and they're working on it is accurate, since it twice told me the wrong time of day. I hope when I wake up in the morning this has all gone away.

 

[Hardus Nameous128 76]

I needed some information about an old building in my town from the 1870's so I asked Google AI about it. A few sentences in to the response I realized it was giving me information about the wrong building; another one down the block. After informing it of its error, confirming the address with Google Maps and re-entering the address it proceeded to apologize for the error and gave me more information for the same incorrect building.

It's getting better, but AI has yet to impress me.

 

[smoke665]

@TJMcG Verizon has had some serious issues with both phone and internet in our state over the last few days that caused cell and internet outages. We use T-mobile so it didn't affect us, however there have been some occasional buffering issues which is unusual.

 

[TJMcG125 8]

@TJMcG Verizon has had some serious issues with both phone and internet in our state over the last few days that caused cell and internet outages. We use T-mobile so it didn't affect us, however there have been some occasional buffering issues which is unusual.

I was not even aware of the major outage that was caused by some kind of software malfunction, but then, yesterday morning, 4 days after the initial outage, I couldn't access some major sites like Reddit, and lesser ones likes Ugly Hedgehog, but Facebook, YouTube, and this one are fine, until you click a link that takes you off-site. Then some work, most don't.

I cannot access Verizon's app or website, but the phone works, so I guess since it's still out this morning, I'll have to call and see what they say. I'm very deaf and it's difficult for me to talk on the phone, even with a Bluetooth hearing aid, but my wife doesn't understand tech talk at all, so it's up to me to get answers. I was very happy with Verizon up to this point, as I had been on Starlink, and this was half the price for faster speeds.

 

[smoke665]

@TJMcG going off thread, but I was a long time customer of Verizon. Almost thru the COVID, DWs phone died. Because of health issues she had to have a phone. Went to the Verizon company store, the door was locked, the only ones inside were employees. Tapped on the glass and a young kid answered, didn't even give us time to explain our situation. Told us we HAD TO have an appointment, and the earliest was a week away. (Again store was empty). Told him where to put his appointment went to their competition walked in, told them our situation, walked out with new service/phones. I'll use smoke signals before I use Verizon again. Customer service stinks and they stick it to you with all their add on fees.

 

[TJMcG125 8]

the ladies at our local Verizon store have been angels. One tried for an hour yesterday to convince the higher-ups that I needed a new router, but they insisted it would work if I just unplugged it one more time. I left there yesterday with instructions to unplug it and plug it back in and it I was sill having problems to come back tomorrow. It went just as you would expect, and I went back this morning where an even more efficient young woman handled them so skillfully that I was in awe. My new router is coming Friday...

 

[AlanKlein94 10]

the ladies at our local Verizon store have been angels. One tried for an hour yesterday to convince the higher-ups that I needed a new router, but they insisted it would work if I just unplugged it one more time. I left there yesterday with instructions to unplug it and plug it back in and it I was sill having problems to come back tomorrow. It went just as you would expect, and I went back this morning where an even more efficient young woman handled them so skillfully that I was in awe. My new router is coming Friday...

I got my own router/modem and have saved a lot of rental money that I was paying for the one Comcast provided me. Does Verizon allow that?