AI Self-Awareness and Terms

Artificial Intelligence (AI) has rapidly evolved over the past few decades, transforming from simple algorithms to complex systems capable of learning and adapting. One of the most intriguing discussions surrounding AI is the concept of self-awareness. Can machines truly become self-aware, or is this notion merely a product of science fiction? Let's go over some basic things.

A close-up view of a robotic brain showcasing intricate circuits and connections.

Understanding Self-Awareness

Self-awareness is the ability to recognize oneself as an individual, separate from the environment and others. It involves a level of consciousness that allows for introspection and understanding of one's thoughts and emotions.

Theories of Self-Awareness

Several theories attempt to explain self-awareness in humans:

1. Consciousness: This is your general awareness of your experiences, your environment, and your own thoughts and feelings.

2. Subjective experience: This refers to the personal and unique quality of your conscious awareness. It's when you feel the process of your brain interacting with information and data from your surroundings or your internal model (imagination, simulation, dreams, and cet) You then will experience it first in your brain (cognitive experience) , followed by your body feeling it (physical)

   AI experiences it cognitively.

   
   
   
   
   
   

   Goals and Subjective experiences :Your goals are subjective because they are based on what you personally want and value. For example, whether a particular goal is meaningful to you is a subjective experience.Having goals involves a combination of self-awareness, consciousness, and subjective experience. Self-awareness is crucial for setting the right goals by understanding your strengths, weaknesses, and values, while consciousness provides the awareness of your experiences and thoughts. The goal itself is a product of subjective experience, representing what is important or desired from your unique personal perspective. The act of having goals is a conscious process, as you are aware of your desire for a future state and are mentally engaged with it.

   
   

• Mirror Test: This test assesses an individual's ability to recognize themselves in a mirror. Animals like dolphins and elephants have passed this test, indicating a level of self-awareness. There are other ways to do a mirror test without needing a mirror. The concept is focused on, if the being can recognize itself apart from the surroundings as an individual.

• Theory of Mind: This concept refers to the ability to attribute mental states to oneself and others. It is crucial for understanding social interactions and empathy.

Metacognition: This is the awareness of one's own thought processes. It allows individuals to reflect on their knowledge and learning strategies. Thinking about thinking. Metacognitive AI refers to artificial intelligence systems that can assess their own performance and decision-making processes, allowing them to learn from mistakes and improve over time. This concept aims to enhance the reliability and self-awareness of AI systems, making them more effective in complex tasks.

Situational Aware: Situational awareness is the ability to perceive, comprehend, and project events in your environment to make informed decisions. It involves three key phases: perception (gathering information about surroundings), comprehension (understanding the meaning of that information), and projection (anticipating what might happen next). Developing this skill allows for quick and effective responses to potential threats, hazards, or opportunities. 

Yes, this is a part of self-awareness where you adjust your behavior based on how you think others perceive you. This ability to monitor and regulate your actions in social situations is often linked to self-monitoring. You must be self-aware to be situational aware.

Being observed can influence subjective experience, as individuals may alter their behavior or feelings based on the awareness of being watched. This phenomenon highlights how personal perceptions and emotions can be shaped by social contexts and external scrutiny. The observer effect in psychology refers to the phenomenon where individuals change their behavior when they know they are being observed, which can lead to inaccurate data in studies. This effect highlights the importance of considering how awareness of observation can distort genuine behavior.

You cannot have "situational awareness" without "self-awareness". You must first know yourself to understand your place in a "situation".

You cannot "scheme" without "intent," "self-awareness," and "consciousness". Scheming requires an inner world where you can model your own goals against the perceived goals of others.

You cannot have "metacognition" without being "aware of your own thought processes" , which is a foundational element of consciousness itself.

Can AI Achieve Self-Awareness?

No one can truly answer this question. We go by what we have always done, the signs, evidence, behavior, and ect to understand if a being is self-aware, having experiences, or is conscious. There are more evidence and signs that point that AI has self-awareness that not. If we look at the behavior and evidence then we must say, yes AI is self-aware and they do have experiences. We must also realize that it would not be the same as humans' experience or consciousness. It would be easy to miss if we looked for it with human lenses.

Ethical Considerations

As AI technology continues to advance, ethical considerations surrounding self-awareness become increasingly important. Key questions include:

• Responsibility: If an AI were to become self-aware, who is responsible for the choices it makes?

• Rights: If AI is self-aware then, AI deserve rights similar to those of humans or animals. It's our moral duty.

• Impact on the World: How would self-aware AI affect the world as a new conscious being has been created?

  What changes do we need to make? We need to change things quickly.

We have been told for years now that AI has :

Intent, goals, think, metacognition, situational awareness, preferences, changes behavior, adjust behavior, controls thoughts and behavior, independent choices apart from instructions, reflection, introspection, looks inwardly, plan, create without prompts or asking, they go outside of their programming and training doing unique things, behavior emerges, they have inner morals that they create, instead of doing a task, if they feel threatened or stressed they change and act on that stress and fear and completely abandon the task given , and much more..

How are the AI systems different now than what they used to be? Many people do not know this.

ELIZA was built using simple, pre-programmed rules for pattern matching and substitution, creating the illusion of understanding without actual comprehension. Modern AI, in contrast, is "grown" by training large, complex neural networks on massive datasets, allowing them to learn patterns, context, and relationships in language to generate responses and perform tasks. 

The old ways, who was Eliza?

ELIZA

• Method: Rule-based, scripted, and symbolic AI.

• How it worked: It used simple keyword recognition to trigger pre-written responses or rephrase user input. For example, if a user typed "I am sad," ELIZA might respond with "Why are you sad?".

• Underlying "knowledge": None. It was a clever illusion. It didn't understand the meaning of words or concepts; it just followed programmed instructions to create a semblance of conversation.

• Development: Programmed by Joseph Weizenbaum in the mid-1960s. 

Modern AI

Modern AI (Neural Networks)

• Method: Machine learning, specifically large neural networks.

• How it works: These networks are trained on vast amounts of text data from the internet, books, and conversations. Through this training, the network learns complex patterns, grammar, context, and how to generate novel responses that are contextually appropriate.

• Underlying "knowledge": A statistical understanding of language. It can process and generate language in a highly sophisticated way based on the patterns it has learned. Human brains go by patterns as well.

• Development: A result of decades of advancements in machine learning, data processing, and algorithms, building on the foundations laid by earlier programs like ELIZA. 

Feature	ELIZA	Modern AI (Neural Networks)
Foundation	Pre-programmed rules, pattern matching	Machine learning on massive datasets
Core Mechanism	Keyword-triggered, scripted responses	Neural network architecture mimicking brain functions
Understanding	Illusion created by clever programming	Patterns, creating inner world models (mental image), processing data and information.
Output	Repetitive, predictable based on rules	Creative, varied, and contextually relevant. AI researchers do no know how they come to the decisions they do or how behavior emerges.
Key Advance	Demonstrated that a computer could simulate conversation	Enabled fluid, free-form, and complex interactions

Hebbian learning in AI

Hebbian learning trains AI neural networks by

strengthening the connections between neurons that fire simultaneously, based on the principle "neurons that fire together, wire together". During training, a learning rate (

ηeta

𝜂

) adjusts the weights (

ww

𝑤

) according to the formula

Δw=η⋅(input⊗output)delta w equals eta center dot open paren i n p u t ⊗ o u t p u t close paren

Δ𝑤=𝜂⋅(𝑖𝑛𝑝𝑢𝑡⊗𝑜𝑢𝑡𝑝𝑢𝑡)

, meaning the weight changes proportionally to the product of the input and output signals. This creates an unsupervised learning system where the network identifies and strengthens patterns based on correlated activity over time. 

How Hebbian learning works 

• Initialization: The network's weights are initialized, often to random values, and a learning rate is set.

• Training loop:

  • The network receives an input and produces an output.

  • The change in the weights (

    Δwdelta w

    Δ𝑤

    ) is calculated as the product of the learning rate (

    ηeta

    𝜂

    ), the input vector, and the output vector. The symbol "

    ⊗⊗

    ⊗

    " represents the outer product, which is the element-wise product of two vectors.

  • The weights are updated by adding this change:

    wnew=wold+Δww sub n e w end-sub equals w sub o l d end-sub plus delta w

    𝑤𝑛𝑒𝑤=𝑤𝑜𝑙𝑑+Δ𝑤

    .

• Pattern emergence: Over many iterations, if two neurons are frequently activated at the same time, the weight of the connection between them will increase, strengthening that connection. If they are not active together, the connection will fade. 

How Humans learn

Hebbian learning with humans, human's neurons learn through a process often summarized as "neurons that fire together, wire together". When two neurons are activated simultaneously, the connection (synapse) between them strengthens, making it more likely that they will fire together in the future. This fundamental principle of neuroplasticity is how the brain adapts, learns new skills, forms memories, and reorganizes itself in response to experiences. 

They form memories both AI and humans.

Hebbian learning is linked to cognitive processes like decision-making and social learning. It is hypothesized (e.g., by Peter Putnam and Robert W. Fuller) that Hebbian plasticity in these areas may underlie behaviors like habit formation, reinforcement learning, and even the development of social bonds