Does Data Know That It Is Data?
Here’s your answer: It does not.
However, there are (hypothetical (for now)) systems that know exactly what they are, and how they came to be. The best example would be humans (who are, in fact, not hypothetical), but let’s think about a field of technology that is much more interesting and rapidly-evolving than humans.
Generative AI (or simply GenAI) refers to a specialised area of artificial intelligence that utilises complex algorithms, particularly deep learning models like Generative Adversarial Networks (GANs) and Variational Autoencoders (VAEs), to learn patterns and structures from vast amounts of data so as to generate new and diverse content that closely resembles the original data, displaying a level of creativity akin to human-made content.
TL;DR: Simply put, generative AI (or GenAI) is a subset of AI systems that is capable of creating original content, data, or information like a human.
While both traditional AI and GenAI use neural networks, the fact that it uses large datasets to generate new content makes GenAI stand out from traditional AI systems, which are usually useful for pattern identification and data analysis.
Even though Generative AI has taken the world by storm, it is not largely different from the AI concepts we are already familiar with; both are built with neural networks, both use the same kind of machine learning techniques, and both systems use insanely large datasets to train, validate, and test their models.
If they’re so similar in nature, why do they yield vastly different results? Traditional AI has applications along the line of pattern recognition and data analysis, but GenAI is at least as good as a human expert when it comes to generating new and original content or data.
At the very least, the answer lies in the architecture and design of the neural networks, and in the structure of the loss functions of the model. They are built specifically for the purpose they are meant to serve, and use different methods of corrections in training to yield the results that are expected, or desired, from them.
GenAI has now given an alarming question (that is as old as the first computer) the stage yet again; what would become of human beings if AI can mimic human intelligence, emotion, and creativity? Is such a thing even possible? Are we going to be invaded by robots?
Here’s the answer to that: we’re more likely to be invaded by aliens than by robots. At least, for now, and for a minimum of a couple more decades.
Super Conscious AI is essentially what all of pop culture thinks robots are. They think and act like (in some works of fiction even resemble) human beings. Then, the question would have to be rephrased as “what would become of human beings if AI can surpass human intelligence, emotion, and creativity”.
In response to this, you might argue that AI models do in fact know what they are, considering this famous phrase: “As an AI language model…”. Sounds pretty self-aware, no?
Here, it is important to realise that it is a model, that has learned what it is (which happens to be an AI language model, if you didn’t already know). It does not inherently know what it is, or know anything for that matter; all of its outputs are results of rigorous training and learning.
Self-aware AI cannot be created (yet) for a multitude of reasons, and at the baseline is the fact that we have, in fact, not defined consciousness yet. As humans, if we don’t know how we know what we are, how can we create something that knows what it is, and also knows how it knows that it knows what it is?
Is that trippy or….?
However, even though we are very far from the development of such an AI system, it is an inevitability to have to consider this question anyway.
Once again, what would become of human beings if AI can mimic human intelligence, emotion, and creativity?
The final answer is: we do not know. It’s more like, we cannot say for sure, but we can predict.
If AI learns to think, feel, and act like a human, isn’t it most likely that the future of AI becomes much like the history of humans? Only, due to the difference of skill level in processing speeds and accuracy, it might be human history in 2x speed. Perhaps it will go through all the kinds of societies, regimes, traditions, conflicts, governances and laws that we have had to go through.
Or, AI may choose not to repeat the same mistakes as humans based on what its learnt by going through all of our history books because of its ability to use logic and ration less discriminately.
Or, since they’re like humans, will they let their experiences and emotions shape who they end up becoming? Will they ponder philosophy sitting under a tree? Can they fall in love? What would death mean to them? Do they become physically enhanced humans?
We won’t know till it happens.
What do you think?
