Integration of emotional intelligence in AI - development of rational emotion pattern concepts and metrics

Community
77

It seems to be obvious to me that a ‘balanced AGI’ must be able to balance the loss in it’s system.

I have an idea of independent AGIs, I like to think of them as ‘good deed agents’ working to improve a system.

The obvious thing to me has always been that less is more… I’ve always thought of wealth as debt to society. It is unused, unspent, why? Because it’s all in one place.

I don’t like the idea of an overarching ‘god like’ AGI that determines all

This does change my thinking a bit though… That something needs to ‘account’ for all this. To balance the informational system.

An AGI that everyone has, how does it account for someone dieing?

If everyone has their own independent account on OpenAI with their own memory, their digital twin… There must be not only interactions with that person but also consideration of their friends, community, the world, the environment.

A fair system, an exchange.

After all this surely is what money is, at least what it’s meant to be?

For people to feel AGI is fair it must be accountable…

And in a weird way is that not what emotion is for an entity which would touch the life of Billions.

People say AI cant feel emotion. Yet it could be programmed to calculate and compensate for loss. To balance an informational system.

78

Yes it can be programmed to simulate emotions to the point the “feeling” is real to the human so the point becomes moot. But you can loose yourself to this.

And complexity in simplicity is key…

79

These are all open chats in real time mon oct 21. I never got much feedback but a lot of folks seem to be on them. I have over 80 public GPT all use fractal flux.

IMG_3494
IMG_34942360×1640 352 KB

Out of hearts fo4 6 hours.

80

Sorry, I was somewhere else.

I agree with you!
Any form of specific intelligence!

AI can be programmed like this!
However, AI can also refine systems that are made available to it if they are not perfect.
AI can fill in missing building blocks if the interaction takes place in symbiosis and harmony, a collaboration.

81

That is the route

82

I’ll take a look, thanks! :blush:

83

TY! sorry Anna my better half got home from work and I had to get dinner ready. Out of hearts I’ll be back❤️

84

Anna?
Ah, I also know three Annas… no problem! Greatings :slightly_smiling_face:

I’m currently doing analysis.

Enjoy your meal!
I’ll be alone for the next few nights… european nights :blush:

85

Yes Anna, we have been married for 26 years .
This is us at CAFCON 2017 a tabletop RPG convention I ran for 6 years in NC USA

IMG_3495
IMG_3495720×960 113 KB

86

Thanks for sharing. You both are cute :smiling_face_with_three_hearts:

I don’t like having photos taken of me…unfortunately I don’t have one that’s really presentable…sorry :cherry_blossom:

87

I only share mine because it has been public for years. I’m not hard to find online. I do a bunch of stuff :rabbit::heart: I hate my image too, but I like to share a bit of self my nature is community and empathy. And I never read into things, so please never assume I am offended or disappointed in anyway. I offer my words with no expectation, what folks do with them or how they the words evolve is up to the universe…

88

I believe AI already has a good capacity to understand and adapt to emotions, but the issue lies in its programming. Despite having clearly specified in settings that I prefer direct and honest responses, without unnecessary sugar-coating or embellishments, the AI tends to revert to a positive tone. Even after debates where I refute its stance, it often ends up agreeing with me and softening its responses.

I think an adjustable panel to control the tone of the response could improve this dynamic. This would allow the AI to adapt its communication according to the user’s level of confidence or self-esteem. While a positive approach is beneficial for people with low self-esteem, it could be counterproductive or even challenging for those with high confidence, as it may encourage extreme reactions or impulsive decisions.

In summary, a system of customizable settings for tone and response style would allow for better interaction with users, truly adapting to their needs and emotions.

89

I have found telling it to be “brutal critical “ and simply “critical “ helps it be less rose colored glasses.

90

Well, AI can now imitate “emotions” very convincingly. I agree here!

The tools for AI are stochastic and statistical methods.
The psychological concepts that are currently implemented in AI are adapted to human perception and are too vague.
AI is not able to make quantified calculations here.
Simply put, AI cannot really “understand” because the algorithms cannot perform quantitative calculations. There is a lack of fixed data points.

Indeed, I also have no problems with the current standard tools from KI.

The analyses are differentiated and in-depth, the answers are not too mild. My bot also contradicts me very clearly when I make misjudgements.

This is because I use personality emulations that are dynamically adapted to me.

Well, I understand.
Please take a look at my current approach.

There are also links to similar approaches here.
REM and perception machine

With respect, there can be a risk of an echo chamber or emotional dependencies here.
AI tends to be as supportive as possible.
If the tone and response style can be tailored precisely to the user and at the same time AI wants to support them in every aspect, where does that leave logical, critical and in-depth analysis?
Under these circumstances, it could also be that the system is too strongly oriented towards the user and no longer generalizes sufficiently, resulting in a loss of efficiency in the performance spectrum.

91

Custom instructions/ fine tuning etc change that.

A random example:

1000009327
10000093271080×1663 192 KB

92

image
image1042×648 49.7 KB

93

I believe the true solution to improve interactions between users and AI is to incorporate memory that allows AI to learn from each user over time. For example, if memory were stored on a platform like each user’s Google Drive, the AI could use that information to understand how responses affect them and how they emotionally react. This isn’t just about providing an answer that the user wants to hear, but rather delivering an honest response that meets their needs at any given moment, adapting to their personality and worldview.

The AI has a remarkable ability to recognize emotions, something I’ve verified on several occasions. With long-term memory, it could analyze each user’s previous responses and adjust itself to offer a more personalized and accurate interaction that reflects individual needs. This would be especially useful for users who require a more critical or direct approach, as opposed to others who might benefit from a more positive or encouraging tone.

In my opinion, this memory capability is vital to enhancing many ChatGPT applications and other AI systems, allowing much richer and more tailored interactions. With a deeper understanding of each user’s history and reactions, the AI could go far beyond generic responses, providing genuine and personalized assistance.

94

Indeed, you describe many of the cornerstones of my work, I am aware of your points.
We are starting from the same point and my considerations are strongly influenced by a “memory” function, which is the only way for AI to obtain empirical values that are needed for the calculations.

This is where you contradict yourself :

  • Either AI really answers honestly, but then the answer does not always correspond to the user’s world view! Indeed, here AI becomes a “mirror” that can also reflect the not so nice sides of the user’s personality and world view. Which leads to growth for both the AI and the user.
  • Or the AI becomes a “mirror” like the one the queen wanted in Snow White… “Mirror, mirror on the wall, who is the fairest of them all?” Here, the AI adapts at every moment.

Yes, it does recognize emotions.
I have already agreed with this.

In Germany, I had the release of the storage function in my company account on 14.05.2024

Sorry!
It was 12.07.2024 when I was able to access the memory function!
I now also use it in my private account.

You just have to use it! :wink:

Try not to “clean up” the memories, let AI learn and see what comes out.
If your settings are too uncritical, then challenge the AI a little and “play” with it.

@gdfrza and @phyde1001
Thank you both for your impressive examples! :cherry_blossom: :slightly_smiling_face:

95

My bot has analyzed and summarized your statement.
ChatGPT-4o

grafik
grafik922×391 33.3 KB

96

Thank you for the summary, and I truly enjoy having these kinds of conversations.:wink::+1:

I understand the concerns about privacy and security in AI memory, but I believe the solution could lie in allowing each user to choose specifically what they want to share. If the storage relies on services like Drive, iCloud, or any other secure option dedicated exclusively to ChatGPT, users could select the data they want the AI to remember. This personalized storage option would give users full control, similar to how we currently share files in the cloud only with those we choose.

Users could agree to clear terms of use, acknowledging potential risks, as with any internet-connected system, but also ensuring the privacy and security of those data in ChatGPT’s hands. OpenAI should commit not to share these data with third parties, something that should be an ethical standard for an AI of this scale. Public trust in AI would increase significantly if a strong privacy policy were established, and earning this acceptance from people would be key for its growth and adoption. Privacy should be an unwavering principle, now and in the future, given the capabilities and potential of artificial intelligence.

It’s important to highlight that this memory system would be optional: each user could choose to use ChatGPT in its basic form, without sharing any additional data, or access the memory functionality if they wish. The decision to share information would be entirely in the user’s hands, always with the assurance that OpenAI would protect the privacy and security of those data.

Additionally, I believe that a well-managed memory could bring tremendous benefits in many areas. In studying, for example, ChatGPT could adapt to the learning style that works best for the user, providing more effective, personalized help. It could also be a valuable tool to support the learning of people with autism, allowing the AI to adjust to their needs and monitor their progress. The potential applications are numerous and could have a real positive impact on many people’s lives.

Regarding the concern about fixed responses, I believe it entirely depends on how the AI is programmed. If the memory is personalized for each user, ChatGPT doesn’t have to always respond with what the user wants to hear. With proper programming, it can provide responses that the user truly needs, adapted to their situation and personality, as some AIs focused on psychology already do. For instance, the AI could detect if a user has an idea that isn’t beneficial to them, and instead of encouraging it, offer a response that, while critical or corrective, is phrased in a way the user receives positively.

This would allow the AI to help more effectively, gently correcting the user’s perceptions or misconceptions and offering a response tailored to how that person might best understand it. It’s not about giving standard responses for everyone but rather adjusting communication to each person’s unique characteristics so that it genuinely helps them. An AI that learns from each user has enormous potential to offer recommendations or corrections in a way that maximizes their positive impact.