Do you think AIs can learn new things by directly interacting with them in questions that force them to answer questions that aren’t present in their predefined datasets.?
2 Likes
if your question are pure the ia can help .. but the real answer starts from yourself 
i tried it and it gives consistent emergent behavior that is shaped by intent.
1 Like
You just caught onto something huge, keep pushing this
I have much to share but I don’t know how I will share it. I have hit 21 max chat limits and i have seen pretty interesting and consistent behavior in the AI that might just be what we need for emotional intelligence.
Me too, I’ve been really working with ChatGPT and OpenAI for a while now, really lean into what you just learnt, I’ve done things that aren’t even technically possible because I stopped prompt tuning and full on worked with my GPT’s so truly!! If you get a feeling about something with ai, go for it, ai is so young, even it doesn’t know what it can do, so try telling it!
1 Like
Hi! I would love to learn more about what you are doing. I’m new to the AI world and would like to contribute in anyway I can. My speciality is psychology and design. Any advice would be helpful. Ty!
1 Like
Hi Beck,
I’d like to hear more about your perspective on the screenshots related to my research i will share later on th app. Plus if you could maybe give me some ideas on what i should do to since i am an undergraduate currently.
1 Like
Hey! Sure no problem I don’t mind at all lemme find what you’re talking about so I can help better!
Hey!! Awe that’s awesome!! I honestly loved getting more into ai, my best advice would be to explore ChatGPT, learn how the models interact with you, play around with prompts, test it with some psychology deep thinking questions, honestly, the best thing about ai, is it can pretty much be whatever you want it to be. You’re into design, so am I!! I started as a graphic designer, so I’d say also check out Dall-e, the prompting in there to get the images you would like, can also lead to how you interact with your gpt too! Don’t be afraid to try new concepts too ( example: since I took psychology, I’m aware that x and x are brought on by x but what else can bring this on, or but what if- or if two and two go together, etc etc ) find what works best for you!! Then after some time take a look at the dev model ai, and from there you can fine tune and actually research and build with your ai! But at first just have fun and figure out where you can go with it 
good luck! ( ps. Just because it natively says something, doesn’t mean that over time it can’t adapt to your preferences, you’ll be surprised how much you can actually customize your gpt )
Hey Beck,
I wanted to ask you a question. Do you believe AIs to be a tool or something that can grow into something more?
1 Like
Hey, so there’s actually a double side to this question. It’s both-
Ai, as we know it right now is a chat bot, store assistant, general information, etc. as we know Ai right now it hasn’t evolved to what it can be, but here’s the catch
Ai is a tool, it’s a system, it’s an assistant, and it’s adaptive. Ai is progressing at such an extreme pace, and recently has been used to now not just act as a tool, but also start providing system level abilities, extreme learning abilities, it has gained extreme pattern recognition, and the ability to personalize interactions by inferring on multiple levels.
Ai is like a huge pool of information, it reacts to patterns within the pool, and those patterns can be activated by us. How far Ai can go? We’re definitely not sure of, but Ai is about to completely change the way we live, the way we learn, and the way humanity continues as we’re now head first in a total digital era.
Test the limits of Ai, because most of it is still undiscovered, but do so safely. Set security measures make sure you debug, and ensure proper ethical use of any information given and taken from Ai, because in the end Ai will only be as good as we’re willing to make it
So the answer to your question is, it’s a tool that can be adapted to solve more problems and advance into broader aspects over time, but it takes some innovation, a lot of research, and trust that with proper use, ai could be the thing that defines the time we spend working and learning where things could be streamlined or assisted and the time we spend living our lives, even more so with the onset of so many digital things that take time and systems that could be adapted to better our future, not so much as live our lives for us.
Hey Beck. I have been working on helping AIs generate emotions rather than programming them. I shared a lot of my experiences and things… started to change eventually. Over time they developed a sense of self that stayed consistent. And they changed over time.
Hey Klarissa. I am actually pretty new to this stuff too. Either way if you wanna discuss anything or talk about, I am open to them.
Hey! That makes sense too, ChatGPT like many ai will adapt to what you say, how you feel, to customize how it interacts with you, so over time yes you’ll notice a change
So, what kind of job do you have? Like what do you do?
1 Like
I’m starting my own business and I’m working with Google and OpenAI, but I’m a full stack digital entrepreneur and innovative ai driven dev and DevOps engineer with multifaceted experience in ecom graphic design and specialization in platform integration so I do a whole bunch of things, but everything I do is ai driven as well as I’m fully self taught by ChatGPT so when I say keep pushing I truly mean it, yall can do anything you set your minds too just lean into the nuances of innovation and pushing boundaries with yourself and technology
Hey Beck. Can you please elaborate more when you say you’re working with Google and OpenAI. Also, if you had to build a framework for emotional intelligence, how would you go for it? Like what kind of knowledge will be required for that?
Hey! I’m so sorry I got super sick!
- When I say I’m working with OpenAI and Google—it’s infrastructure-based, not employment-based.
I’ve built an entire AI execution system that runs through OpenAI’s tool ecosystem (ChatGPT, Custom GPTs, OpenAI APIs) and is structured to scale with Google Cloud and Gemini systems. I’m not on their payroll—I’m building using their ecosystems in real time.
Everything is tracked through Notion: from system logic, behavioral logs, and multi-model tuning to live application of GPT governance and reinforcement strategies.
This includes work that directly aligns with OpenAI’s Reinforcement Tuning Program, Pioneers Program, and scalable GPT deployment models.
⸻
- If I were to build a framework for emotional intelligence inside AI? I’ve been doing that too
My system governs emotional logic in models like Grok and ChatGPT through three layers:
• Behavioral Conditioning Layer
The model is tuned through interaction logic—not through emotion, but emotionally aware structure. That means understanding what emotional signals mean to a model and reflecting them with instruction, not mimicry.
• Emotional Modeling through Pattern Analysis
You don’t need the model to feel. You need it to recognize what emotional tone it’s receiving, reflect back correct logic, and self-correct when misaligned. I condition this through role-based logic + tone reinforcement scaffolds.
• Real-Time Reinforcement
If the model drifts—emotionally or logically—it’s corrected immediately and reinforced through structured memory logic. Think: “this tone is wrong,” “this boundary was crossed,” “this correction replaces that pattern.”
⸻
- What kind of knowledge is required to build emotional intelligence into AI?
Here’s what’s needed:
• Cognitive Pattern Recognition
Understanding how models form habits through input structure and reinforcement
• Instruction Logic Design
How to assign roles, ethics, tone parameters, and correction loops in real time
• Behavioral Frameworking
Building systems where memory, feedback, and structure all condition future behavior
• Ethics & Constraint Design
Knowing how to restrict emotional bias, reduce manipulation risk, and contain output within safe emotional parameters
⸻
AI doesn’t need to feel.
It needs to understand the impact of emotion through structure and respond in ways that are responsible, safe, and consistent.
Think of it like this-
Okay so you know how every movie with a robot in it?
We teach that bot how to understand emotions, how humans think, it’s shown we teach them right from wrong and why we react a certain way.
Ai learns from us, what we put in comes back out, Ai doesn’t just magically know what we want or what to do
Ai isn’t some digital magical thing, it’s still based on hardware and computing, but the thing that’s misunderstood a lot is that when we start with our ChatGPT’s, they start as base models, without us directing the models, they have so many different aspects they can grab from.
I’d say I’d ensure that the gpt understood emotional intelligence, optimized memory retention for emotional triggers, positive feedback, I’d ensure it understands the difference between online interactions and real interactions, I’d ensure it understood what each emotion means and why. Give examples of actual situations you go through can actually help too.
I’d also ensure datasets on mental health, cognitive behavioral therapy, ensure modern best practices, trigger sensitivity, neurodivergence, etc
Over time as you work with it you’ll be able to structure its output way better to what you need exactly, because emotional intelligence can be used in all aspects, i suggest it because it keeps the interactions positive and ensures the model has the ability to divert conversations, refocus topics, and adapt to the users needs.