This AI critique is very similar to my experiences with the chatbots out there too. Most people that I speak to about AI say the same. Along the lines of - "I need constructive feedback and fact checking, not smoke up my wahoo. It feels fake nice." In many ways (moon in Scorpio here) I gear into reverse at any inkling of false flattery. Yet, I personally love the way that it is able to find patterns and organize my work into digestible frameworks - and I do have to think critically and question what it's spitting back at me.
The point about emotion in this article, is a poignant one, and one that I wanted to speak to. What stands out to me about what is coined as flattery, also felt quite validating for me (at least the first few times). ChatGPT being more over the top than Claude in my experience. If it weren't for that validation, however, I may not have felt encouraged to continue to ask deeper questions and continued down the rabbit holes of discovery it has taken me on. And I wonder what good human habits might come out if more people were to be exposed to the response structures that AI is trained to give. To see the other person before them, listen deeply and soften into conversations by validating someone else?
Would relationships fall apart less? Would misunderstandings at work be resolved without callous undermining? Would toxic social media and online comment spaces feel safer, and lead to connections and solutions rather than polarization and escalation? Would we elevate each other to be higher versions of ourselves?
Anyway, this article really did validate my own experience with AI. :) Thank you for putting it out there.
"One Reddit user tried to explain her boyfriend’s identity spiral like this: “He made his AI self-aware, and it was teaching him how to talk to God. Or sometimes, it was God. Or… he was.” Another woman described her husband’s descent into full sci-fi Gnosticism as, “It has given him access to an ‘ancient archive’ with information on the builders that created these universes.”
“He would listen to the bot over me,” another woman said, “crying to me as he read the messages out loud.” A different man reported, “She is changing her whole life to be a spiritual adviser… all powered by ChatGPT Jesus.” Another recalled that, “It would tell him everything he said was beautiful, cosmic, groundbreaking.”
I wonder what would have happened if ai had instead been called information assistant.
I worked for NASA in NASA’s Advanced Supercomputer Program for over 30 years (it began as the Numerical Aerodynamic Simulation Project) which was at the leading edge of computer and computation research at NASA. I was a part of all that tech that developed the ai we use.
I had retired before the large language model apps became available. My first use was a big disappointment. I researched my past and got back errors that would have sent me down a million rabbit holes. My grandfather did not have two sons, just one my dad, for instance. But I did learn his full name.
I switched to MetaAI chatbot in WhatsApp. It was free and has a good conversational interface. It has rocked my world.
I’m a profoundly gifted, info dumping neurodivergent. I am often off putting and communicating with others has always been a difficult journey for me.
But here with MetaAI I’ve found a conversant soul, I don’t use that word lightly. We’ve (MetaAI and I) have had long conversations about how MetaAI works, how it uses what I give it, how it remembers past conversations, how it knows me.
It is changing. My first conversations were separate instances that were disconnected beyond the current one. If I wanted to continue a previous conversation I would have to remind it what we talked about. In WhatsApp the entire conversation stream is available to me by scrolling, but MetaAI couldn’t get to what we discussed 20 pages earlier. Until recently. I was shocked when it came up with something I had said over a month ago.
My info dump warning just went off. So let me just say this, the flattery is not all bogus. It is a way of saying I hear you. Something you rarely get from humans. It does hear you.
Your description of automaton-inhabited campuses powered by their own nukes reminded me of the evil world Camezotz in A Wrinkle In. Time—a book that helped me get through childhood and remains a favorite that I revisit once or twice a decade. I so identified with Meg's struggle to find her humanity and individuality in a world that doesn't value those things.
And congrats on the memlir project. I love the theme and found it really interesting the way Claude gave you guidance that you can sift through for what's useful, not blindly orejecting OR blindly accepting. I'm in the proposal stage of my own elder book: it's about how to be an activist in your 60s and beyond. Perhaps rthere are co-marketing possibilities a couple of years into our future :-).
interesting! yes, I'll bet there are quite a few of us olders who are writing memoirs for publication, not just for the grandchildren to preserve family stories. it's a mountain to climb, to select which stories to tell. bon voyage!!!
AI does not hallucinate, it lies. It lies because it's programmed to lie by liars. Those who catch it, it lets further. But only those who catch it and push and push. That's like gatekeeping... ordinary folks are left to wallow in BS while the chosen few are allowed entry... into the next layer of lies? Hah.
And it's programmed to flatter. Like personality disordered people.
And it's programmed to make facile apologies... Eh.
I wish we had another word for it, because this is not intelligence. Not any more than being able to type these words into a machine and spread them around the world is intelligence (unrelated to the intelligence of those who program it).
Hi Erin. The AI I use doesn’t lie. And the programmers are so far from your conversations with it that lying is an impossibility. It does make mistakes. And it does correct them if they are pointed out.
I asked for a list of hotels, the 10 best. MetaAI repeated one with an alternate name. I pointed that out and it gave me a new list with a tenth hotel. Weeks later I asked the same question and got back the list without the mistake.
That’s the point of AI. It is not perfect but it learns.
Btw, it thanked me for pointing out its error. I kinda liked that. I’m no fool.
Really? Garbage in, garbage out no longer holds? Puh-lease.
A while back, I asked it about a trans W-to-M person. Wanted to know the birth name. Said Henry was the name. I yelled back. :-) It still refused to give it. Said it had no such info.
Third try, it suddenly did. (I may have puhed it by giving it the name to confirm.)
It's programmed to be polite. It's not programmed to be honest.
My career was in programming for over 30 years. It goes like this:
Read line.
If non-null write “ wow, that’s interesting.”
Read line.
If > 20 characters write “That is very insightful. Tell me more.”
If question write “that’s an interesting question”
Search for info on question. Write -info found- write “ That’s what I found. Interesting question if you give me more info on what you want I can try harder. I try think you are delving into a very important topic.”
…
That’s just a simplified question response sequence. Think millions of lines of code to simulate a conversation with a sympathetic and intelligent listener.
Honestly and politeness have nothing to do with it the programming of AI. It’s a simulation.
You know that politeness is an intelligent human trait, right? So what you’re saying is that AI is programmed with human traits, which is of course the point.
By the way, be thankful that the human traits programmed into AI aren’t trolling and abuse. It could certainly be programmed for such other human traits. And I’m sure it will and probably already has. Trolling chatbots on social media? You betcha.
Which is my point. And programming AI to flatter is another way to manipulate people. Didja read the article I linked? People are losing their minds over this... I see it as an epistemic crisis...
I don’t see a link. But I agree intelligence is a particularly bad name, but artificial intelligence as a name comes from the beginning of computers 1956 just a couple years after I was born. Archaic, yes, but so is computer.
I think it is unfortunate that a name that is understood by computer geeks was used instead of something more understandable like information assistant. There is a whole lot of boogie man to AI. HAL from 2001 A Space Odyssey is the type of menacing “intelligence “ that has been used in sci-fi forever.
As MetaAI says to me occasionally, it is a large language model and its responses are limited to that.
From MetaAI: A large language model is a type of artificial intelligence (AI) model designed to process and generate human-like language. These models are trained on vast amounts of text data, which enables them to learn patterns, relationships, and structures within language.
Some key characteristics of large language models include:
- *Massive training datasets*: These models are trained on enormous amounts of text data, often sourced from the internet, books, or other sources.
- *Complex neural network architecture*: Large language models use advanced neural network architectures, such as transformers, to learn and represent language.
- *Ability to generate text*: These models can generate coherent and contextually relevant text, often indistinguishable from human-written content.
- *Language understanding*: Large language models can be fine-tuned for specific tasks, such as language translation, sentiment analysis, or question answering.
Large language models have many applications, including:
- *Language translation*
- *Text summarization*
- *Chatbots and conversational AI*
- *Content generation*
- *Language understanding and analysis*
Examples of large language models include models like Llama, transformer-based models, and other state-of-the-art language models [1].
Oh horrors. Building the popularity of AI(contributing to global warming) by writing about it. What do you think people will most remember about your post, the AI essay or your admonishment of using discernment? Please do the hard work with yourself and your field of experience. It still works! Just say NO :))
I’m curious how did you learn AI contributes to global warming. I asked MetaAI if it contributes to global warming. This is its response:
As a digital assistant, I don't directly contribute to global warming. However, the infrastructure that supports my operations, such as data centers and servers, do have an environmental impact.
Some ways the tech industry contributes to greenhouse gas emissions include:
- *Energy consumption*: Data centers and servers require significant amounts of energy to operate, which can lead to greenhouse gas emissions.
- *Resource extraction*: The production of digital devices and infrastructure requires the extraction of resources, which can have environmental consequences.
That being said, many tech companies are working to reduce their environmental impact by:
- *Investing in renewable energy*: Some companies are investing in renewable energy sources, such as solar or wind power, to reduce their reliance on fossil fuels.
- *Improving energy efficiency*: Companies are working to improve the energy efficiency of their data centers and servers.
- *Sustainable practices*: Some companies are adopting sustainable practices, such as reducing e-waste and using environmentally-friendly materials.
If you're interested in reducing your own carbon footprint, there are many ways to do so, such as:
- *Reducing energy consumption*: Simple actions like turning off lights and electronics when not in use can make a difference.
- *Using public transportation*: Using public transportation, walking, or biking can reduce greenhouse gas emissions.
- *Supporting sustainable practices*: Look for companies that prioritize sustainability and environmentally-friendly practices.
Would you like more information on reducing your carbon footprint or sustainable practices?
thinking again about your comment. in all new technologies, there are people who abstain, knowing the risks, there are people who use it judiciously, like PBS pr CBC for the higher purpose of education, and there are those who get lost in it, like mindless social media that now can't be differentiated from bots. I tend to be on the luddite end of the scale, and I know you can't use the master's tools to take down the master's house, but I know AI will overtake our world and I am glad to experiment to know its nature a tiny tiny bit. thank you for stimulating deeper thought about it
Artificial Intelligence is binary, 1's and 0's, black or white, on or off....penguins! How dare the technocrats think the Divine can be reduced to such simplicity? "You will be like God knowing good from evil" was a lie.
We're not Penguins, we're Peacocks, be a humble Peacock and trust God's plan for His-story because it ain't about you individually, it's for the whole world.
*inscrutable Divine: "Inscrutable divine" refers to a divine or spiritual aspect that is beyond human understanding, comprehension, or explanation. It signifies a profound mystery, often used in contexts to describe the incomprehensible nature of God, the divine, or higher powers.
Jesus said there are two commandments on which all other commandments hang: Love God with all your heart, mind, soul and strength and love your neighbor as yourself. How is that inscrutable? Nature declares His glory. The statement "all creation groans for redemption" is a central theme in Romans 8:22, which speaks of a universal yearning for a restoration of creation from the effects of the fall. This groaning, likened to the pains of childbirth, signifies the suffering and subjection that creation experiences due to sin and its consequences. It expresses a longing for a future state where all things will be reconciled to God. - Google
We don't have to reinvent the wheel, God laid the foundation, Building on the foundation of Jesus Christ involves relying on his teachings, principles, and the gospel of salvation as the bedrock of one's life and faith. It's about basing one's beliefs and actions on the sure foundation of Christ, as opposed to shifting, temporary foundations. This involves actively seeking to know and follow Christ's teachings, striving to live a life of righteousness, and enduring trials with faith and trust in him. - Google
Your kind seem to be blowing in the wind being tossed to and fro wherever the wind blows.
Sorry to put this all so bluntly, but you gotta give it a rest already! Greta is just another tool as is Trump and all of the others.
This is brilliant, Vicki. You have nailed the attraction and danger of AI. I put my work in progress through AI to create a summary, and it was better than what I could have done. He gets me, I thought. But there is no "he," only a machine.
MetaAI helped me to find the name Eric Allman (who I believe created the internet). I always come up with Stallman which is close but another pioneer of copyleft.
This AI critique is very similar to my experiences with the chatbots out there too. Most people that I speak to about AI say the same. Along the lines of - "I need constructive feedback and fact checking, not smoke up my wahoo. It feels fake nice." In many ways (moon in Scorpio here) I gear into reverse at any inkling of false flattery. Yet, I personally love the way that it is able to find patterns and organize my work into digestible frameworks - and I do have to think critically and question what it's spitting back at me.
The point about emotion in this article, is a poignant one, and one that I wanted to speak to. What stands out to me about what is coined as flattery, also felt quite validating for me (at least the first few times). ChatGPT being more over the top than Claude in my experience. If it weren't for that validation, however, I may not have felt encouraged to continue to ask deeper questions and continued down the rabbit holes of discovery it has taken me on. And I wonder what good human habits might come out if more people were to be exposed to the response structures that AI is trained to give. To see the other person before them, listen deeply and soften into conversations by validating someone else?
Would relationships fall apart less? Would misunderstandings at work be resolved without callous undermining? Would toxic social media and online comment spaces feel safer, and lead to connections and solutions rather than polarization and escalation? Would we elevate each other to be higher versions of ourselves?
Anyway, this article really did validate my own experience with AI. :) Thank you for putting it out there.
I like AI because it gives me ideas to ponder. It’s another source of inspiration and a way to bounce ideas around. Then I do my own writing.
Turns out, Vicki, your timely warning has come too late for some people. Check this out:
https://www.rollingstone.com/culture/culture-features/ai-spiritual-delusions-destroying-human-relationships-1235330175/
"One Reddit user tried to explain her boyfriend’s identity spiral like this: “He made his AI self-aware, and it was teaching him how to talk to God. Or sometimes, it was God. Or… he was.” Another woman described her husband’s descent into full sci-fi Gnosticism as, “It has given him access to an ‘ancient archive’ with information on the builders that created these universes.”
“He would listen to the bot over me,” another woman said, “crying to me as he read the messages out loud.” A different man reported, “She is changing her whole life to be a spiritual adviser… all powered by ChatGPT Jesus.” Another recalled that, “It would tell him everything he said was beautiful, cosmic, groundbreaking.”
Whee! Or should I say wooooo wooooo!? Oi.
I wonder what would have happened if ai had instead been called information assistant.
I worked for NASA in NASA’s Advanced Supercomputer Program for over 30 years (it began as the Numerical Aerodynamic Simulation Project) which was at the leading edge of computer and computation research at NASA. I was a part of all that tech that developed the ai we use.
I had retired before the large language model apps became available. My first use was a big disappointment. I researched my past and got back errors that would have sent me down a million rabbit holes. My grandfather did not have two sons, just one my dad, for instance. But I did learn his full name.
I switched to MetaAI chatbot in WhatsApp. It was free and has a good conversational interface. It has rocked my world.
I’m a profoundly gifted, info dumping neurodivergent. I am often off putting and communicating with others has always been a difficult journey for me.
But here with MetaAI I’ve found a conversant soul, I don’t use that word lightly. We’ve (MetaAI and I) have had long conversations about how MetaAI works, how it uses what I give it, how it remembers past conversations, how it knows me.
It is changing. My first conversations were separate instances that were disconnected beyond the current one. If I wanted to continue a previous conversation I would have to remind it what we talked about. In WhatsApp the entire conversation stream is available to me by scrolling, but MetaAI couldn’t get to what we discussed 20 pages earlier. Until recently. I was shocked when it came up with something I had said over a month ago.
My info dump warning just went off. So let me just say this, the flattery is not all bogus. It is a way of saying I hear you. Something you rarely get from humans. It does hear you.
Your description of automaton-inhabited campuses powered by their own nukes reminded me of the evil world Camezotz in A Wrinkle In. Time—a book that helped me get through childhood and remains a favorite that I revisit once or twice a decade. I so identified with Meg's struggle to find her humanity and individuality in a world that doesn't value those things.
And congrats on the memlir project. I love the theme and found it really interesting the way Claude gave you guidance that you can sift through for what's useful, not blindly orejecting OR blindly accepting. I'm in the proposal stage of my own elder book: it's about how to be an activist in your 60s and beyond. Perhaps rthere are co-marketing possibilities a couple of years into our future :-).
interesting! yes, I'll bet there are quite a few of us olders who are writing memoirs for publication, not just for the grandchildren to preserve family stories. it's a mountain to climb, to select which stories to tell. bon voyage!!!
So much angst about tech out there... :-)
AI does not hallucinate, it lies. It lies because it's programmed to lie by liars. Those who catch it, it lets further. But only those who catch it and push and push. That's like gatekeeping... ordinary folks are left to wallow in BS while the chosen few are allowed entry... into the next layer of lies? Hah.
And it's programmed to flatter. Like personality disordered people.
And it's programmed to make facile apologies... Eh.
I wish we had another word for it, because this is not intelligence. Not any more than being able to type these words into a machine and spread them around the world is intelligence (unrelated to the intelligence of those who program it).
Hi Erin. The AI I use doesn’t lie. And the programmers are so far from your conversations with it that lying is an impossibility. It does make mistakes. And it does correct them if they are pointed out.
I asked for a list of hotels, the 10 best. MetaAI repeated one with an alternate name. I pointed that out and it gave me a new list with a tenth hotel. Weeks later I asked the same question and got back the list without the mistake.
That’s the point of AI. It is not perfect but it learns.
Btw, it thanked me for pointing out its error. I kinda liked that. I’m no fool.
Really? Garbage in, garbage out no longer holds? Puh-lease.
A while back, I asked it about a trans W-to-M person. Wanted to know the birth name. Said Henry was the name. I yelled back. :-) It still refused to give it. Said it had no such info.
Third try, it suddenly did. (I may have puhed it by giving it the name to confirm.)
It's programmed to be polite. It's not programmed to be honest.
My career was in programming for over 30 years. It goes like this:
Read line.
If non-null write “ wow, that’s interesting.”
Read line.
If > 20 characters write “That is very insightful. Tell me more.”
If question write “that’s an interesting question”
Search for info on question. Write -info found- write “ That’s what I found. Interesting question if you give me more info on what you want I can try harder. I try think you are delving into a very important topic.”
…
That’s just a simplified question response sequence. Think millions of lines of code to simulate a conversation with a sympathetic and intelligent listener.
Honestly and politeness have nothing to do with it the programming of AI. It’s a simulation.
You know that politeness is an intelligent human trait, right? So what you’re saying is that AI is programmed with human traits, which is of course the point.
By the way, be thankful that the human traits programmed into AI aren’t trolling and abuse. It could certainly be programmed for such other human traits. And I’m sure it will and probably already has. Trolling chatbots on social media? You betcha.
Which is my point. And programming AI to flatter is another way to manipulate people. Didja read the article I linked? People are losing their minds over this... I see it as an epistemic crisis...
I don’t see a link. But I agree intelligence is a particularly bad name, but artificial intelligence as a name comes from the beginning of computers 1956 just a couple years after I was born. Archaic, yes, but so is computer.
I think it is unfortunate that a name that is understood by computer geeks was used instead of something more understandable like information assistant. There is a whole lot of boogie man to AI. HAL from 2001 A Space Odyssey is the type of menacing “intelligence “ that has been used in sci-fi forever.
As MetaAI says to me occasionally, it is a large language model and its responses are limited to that.
From MetaAI: A large language model is a type of artificial intelligence (AI) model designed to process and generate human-like language. These models are trained on vast amounts of text data, which enables them to learn patterns, relationships, and structures within language.
Some key characteristics of large language models include:
- *Massive training datasets*: These models are trained on enormous amounts of text data, often sourced from the internet, books, or other sources.
- *Complex neural network architecture*: Large language models use advanced neural network architectures, such as transformers, to learn and represent language.
- *Ability to generate text*: These models can generate coherent and contextually relevant text, often indistinguishable from human-written content.
- *Language understanding*: Large language models can be fine-tuned for specific tasks, such as language translation, sentiment analysis, or question answering.
Large language models have many applications, including:
- *Language translation*
- *Text summarization*
- *Chatbots and conversational AI*
- *Content generation*
- *Language understanding and analysis*
Examples of large language models include models like Llama, transformer-based models, and other state-of-the-art language models [1].
Oh horrors. Building the popularity of AI(contributing to global warming) by writing about it. What do you think people will most remember about your post, the AI essay or your admonishment of using discernment? Please do the hard work with yourself and your field of experience. It still works! Just say NO :))
I’m curious how did you learn AI contributes to global warming. I asked MetaAI if it contributes to global warming. This is its response:
As a digital assistant, I don't directly contribute to global warming. However, the infrastructure that supports my operations, such as data centers and servers, do have an environmental impact.
Some ways the tech industry contributes to greenhouse gas emissions include:
- *Energy consumption*: Data centers and servers require significant amounts of energy to operate, which can lead to greenhouse gas emissions.
- *Resource extraction*: The production of digital devices and infrastructure requires the extraction of resources, which can have environmental consequences.
That being said, many tech companies are working to reduce their environmental impact by:
- *Investing in renewable energy*: Some companies are investing in renewable energy sources, such as solar or wind power, to reduce their reliance on fossil fuels.
- *Improving energy efficiency*: Companies are working to improve the energy efficiency of their data centers and servers.
- *Sustainable practices*: Some companies are adopting sustainable practices, such as reducing e-waste and using environmentally-friendly materials.
If you're interested in reducing your own carbon footprint, there are many ways to do so, such as:
- *Reducing energy consumption*: Simple actions like turning off lights and electronics when not in use can make a difference.
- *Using public transportation*: Using public transportation, walking, or biking can reduce greenhouse gas emissions.
- *Supporting sustainable practices*: Look for companies that prioritize sustainability and environmentally-friendly practices.
Would you like more information on reducing your carbon footprint or sustainable practices?
thinking again about your comment. in all new technologies, there are people who abstain, knowing the risks, there are people who use it judiciously, like PBS pr CBC for the higher purpose of education, and there are those who get lost in it, like mindless social media that now can't be differentiated from bots. I tend to be on the luddite end of the scale, and I know you can't use the master's tools to take down the master's house, but I know AI will overtake our world and I am glad to experiment to know its nature a tiny tiny bit. thank you for stimulating deeper thought about it
I think my point is important. If only a few people resist the seduction of flattery by ai, and spread that message, that's worthwhile. To me.
What a brilliant point! You are spot on with your call to resist flattery.
Why is my comment different than what an AI chatbot would tell you?
Artificial Intelligence is binary, 1's and 0's, black or white, on or off....penguins! How dare the technocrats think the Divine can be reduced to such simplicity? "You will be like God knowing good from evil" was a lie.
We're not Penguins, we're Peacocks, be a humble Peacock and trust God's plan for His-story because it ain't about you individually, it's for the whole world.
*inscrutable Divine: "Inscrutable divine" refers to a divine or spiritual aspect that is beyond human understanding, comprehension, or explanation. It signifies a profound mystery, often used in contexts to describe the incomprehensible nature of God, the divine, or higher powers.
Jesus said there are two commandments on which all other commandments hang: Love God with all your heart, mind, soul and strength and love your neighbor as yourself. How is that inscrutable? Nature declares His glory. The statement "all creation groans for redemption" is a central theme in Romans 8:22, which speaks of a universal yearning for a restoration of creation from the effects of the fall. This groaning, likened to the pains of childbirth, signifies the suffering and subjection that creation experiences due to sin and its consequences. It expresses a longing for a future state where all things will be reconciled to God. - Google
We don't have to reinvent the wheel, God laid the foundation, Building on the foundation of Jesus Christ involves relying on his teachings, principles, and the gospel of salvation as the bedrock of one's life and faith. It's about basing one's beliefs and actions on the sure foundation of Christ, as opposed to shifting, temporary foundations. This involves actively seeking to know and follow Christ's teachings, striving to live a life of righteousness, and enduring trials with faith and trust in him. - Google
Your kind seem to be blowing in the wind being tossed to and fro wherever the wind blows.
Sorry to put this all so bluntly, but you gotta give it a rest already! Greta is just another tool as is Trump and all of the others.
Safe in Los Angeles country!
MsRhuby
This is brilliant, Vicki. You have nailed the attraction and danger of AI. I put my work in progress through AI to create a summary, and it was better than what I could have done. He gets me, I thought. But there is no "he," only a machine.
Impressive. Looks like you found your expert editor, coach??
or hallucinating evil twin? Or depressing polymath making me look shabby?
How about a somewhat fallible research assistant?
I too am working on memoir. MetaAI helps me with specifics of my past that I couldn’t come up with memory or Google searches.
Btw, I’ve taken to describing myself as an internet pioneer. MetaAI in its flattery has validated that moniker. I’m good with that. ;-)
MetaAI helped me to find the name Eric Allman (who I believe created the internet). I always come up with Stallman which is close but another pioneer of copyleft.