In 2024, AI is just an unconscious gifted child

Imagine a world where Artificial Intelligence (AI) takes on a human form. If AI was a human being in 2024, it would not yet be an adult, nor even a teenager. Instead, AI would probably be a gifted child — full of potential, impressive in its capabilities, but still learning and growing, with a long road ahead before reaching full maturity.?

I was searching for the right image to explain the maturity level of AI to a friend. And I thought about an HPI/gifted child : capable of remarkable things, yet fundamentally immature in its understanding and application of the world around it.

Like a gifted child who quickly picks up new skills and knowledge, AI has developed rapidly in recent years. The technological leaps have been extraordinary, like a young prodigy mastering complex concepts that even adults find challenging. For that, the genius kid is using his/her phenomenal memory and a fantastic ability to learn and improve over time.

AI systems today can understand and generate human language with surprising fluency, play strategy games like chess and Go at superhuman levels, and even create art and music that stir human emotions. These achievements are reminiscent of a child who, at a young age, demonstrates a talent for music by playing a piano sonata or solving advanced mathematical problems. Just as such a child captivates and impresses those around them, AI's feats have captured the imagination of the world, sparking both awe and curiosity about what it might achieve next.

Human vs AI Brains?

Yet, despite these capabilities, AI remains like a child in many crucial ways. A gifted child may be able to solve a Rubik's Cube in seconds or recite Shakespeare from memory, but they still lack the deep understanding and experience of an adult. The emotional intelligence needed? to connect with our hearts, not only our brains.

Brains…when I started working on AI landscape, I heard a lot about neural networks and transformers. So quite logically,? I was interested in comparing the human brain and the current state of “AI brain”.

The human brain consists of about 86 billion neurons interconnected by trillions of synapses. It operates in a highly parallel and distributed manner, with different brain regions specialising in various tasks. The brain's structure allows it to process information from sensory inputs, generate thoughts, and control bodily functions simultaneously.

A Transformer model, like GPT-4, is a type of artificial neural network designed for processing sequences, particularly text. It is composed of layers of attention mechanisms and feedforward neural networks. The Transformer operates sequentially (in layers) rather than in the parallel, distributed way the brain does. Its architecture is built to handle large amounts of data with a focus on relationships within sequences, like words in a sentence.

According to a Deloitte report, “ChatGPT? was initially trained with more than 1 million datasets or 500 billion tokens (words or word fragments), including from Wikipedia and The New York Times. To put this into perspective, the average human speaks 860.3 million words in their lifetime, making this collection — or “corpus,” in AI terminology — equivalent to 300 years’ worth of language”

In a nutshell : AI Brain is way bigger than Human Brain. But the magical complexity of our human brain (and heart) is almost impossible to clone for machines and algorithms.

Divine consciousness?

Why ? I tried to find the answer in Descartes' statement : “Cogito ergo sum, I think therefore I am”.

Consciousness and awareness.

Consciousness is one of the most profound aspects of the human brain, encompassing self-awareness, the ability to reflect, and subjective experiences. Humans are capable of introspection, understanding their own thoughts, and experiencing emotions in a way that is deeply tied to their physical and social existence. AI, including Transformer models, has no consciousness, awareness, or subjective experience. It processes data and generates outputs without understanding or meaning. Transformers do not have desires, motivations, or emotions, their "outputs" are devoid of any personal experience. It’s a neural network not an emotional system.

One of the most profound challenges is the "hard problem" of consciousness, which refers to explaining how subjective experience arises from physical processes. This problem remains unsolved in neuroscience and philosophy, and it poses a significant barrier to the idea of AI consciousness. But enough with intellectual peregrination..

To make it simple, while an AI model like GPT-4 can generate a convincing essay or hold a conversation, it does not truly understand the text in the way a human does. It doesn't grasp the nuances, emotions, or contexts behind the words it uses. It’s like a child who, though able to speak in full sentences, doesn’t yet fully understand the subtleties and complexities of adult communication.

A child learns from experiences, stories, and emotions, drawing connections across various aspects of life, from discussions and fights with parents trying to explain the fundamental differences between bad and good. A child observes, listens and forges his or her character based on emotional and intellectual figures of authority.

In contrast, AI learns from massive amounts of data, finding patterns and correlations without any real comprehension of the underlying meaning.?

Le cancre?

When I was a child, my mum forced me to learn “Le Cancre” from Jacques PREVERT. She was? obsessed with this poem : “Il dit non avec la tête mais il dit oui avec le c?ur, il dit oui à ce qu’il aime, il dit non au professeur”.?

I am 41 years old and I remember every single line since I was 9 or 10 years old. I am 41 years old, and I really started understanding the meaning of this poem when I was 20 (thanks mum).?

This poem is about freedom, listening and following your heart, without being worried about normality and social rules.? I knew the words, but I had no clue about the meaning of this beautiful text and why it was so important for my mother.

A gifted child often knows a lot about a few topics but may struggle outside those areas. Similarly, AI systems are typically designed to excel in narrow domains — such as recognizing faces, playing a game, or predicting consumer behaviour.?

Just so you know : we call it “weak AI” (or narrow AI).

These systems are powerful within their specific boundaries but are lost when it comes to perform a totally new task outside of their core realm.Imagine asking a child who is a chess prodigy to suddenly perform a complex dance routine. The child might not know where to begin. Likewise, an AI trained to identify tumours in medical images will be lost if asked to write a poem.?

The future?

Despite these limitations, we should not underestimate AI’s potential, just as we should never underestimate the potential of a gifted child. The rapid pace of AI research and development suggests that we are merely at the beginning of this journey. Everyday, AI systems become more sophisticated, learning to perform new tasks, adapting to new environments, and getting better at mimicking human thought processes. In the coming years, AI will become a teenager — still not fully mature but with a broader understanding of the world and a more versatile set of skills.?

However, the dream (or nightmare) of reaching Artificial General Intelligence (AGI), where machines can understand, learn, and apply knowledge across a wide range of tasks as humans do, remains a VERY hypothetical.?

Like raising a gifted child, guiding the development of AI requires careful nurturing, ethical considerations, and a focus on long-term well-being. As AI systems become more integrated into our daily lives, it is crucial to ensure they are developed and used responsibly. We must teach this "child" not only to be smart but also to be wise — to understand fairness, ethics, and the implications of its actions.

This includes addressing issues like bias in AI algorithms and datasets, ensuring transparency in decision-making processes, and protecting user privacy. Just as a child needs guidance to grow into a responsible adult, AI needs oversight and thoughtful development to ensure it benefits humanity as a whole. Exciting time ahead…

Nicolas Votano (inspired by ChatGPT;) or ChatGPT (inspired by Nicolas Votano) - your choice.