ChatGPT is the most advanced language model that exists today, it works by using a deep learning algorithm known as a Transformer. This algorithm has been trained on a massive amount of data from the internet, including text from books, articles, and websites. As a result, it has the ability to generate human-like text based on the input it receives. It is terryfying how humanlike it's reponses are. Try for yourself.
Apart from the fact that it'll be an upgare from google search in terms of helping lazy students solve their homeworks with a click, it has serious dark potential.
Language models and transformers.
First, let's quickly review what these models are and how they work!
Language models are a type of artificial intelligence that are trained to generate text that resembles human language. They are used in a variety of applications, such as machine translation, text generation, and text classification.
A Transformer is a specific type of language model that has been designed to process sequential data, such as text. It was introduced in the paper "Attention is All You Need" by Vaswani et al. in 2017 and has since become the dominant architecture for natural language processing tasks.
Language models, including Transformers, are trained using large amounts of text data, typically from the internet. During the training process, the model is exposed to the text data and adjusts its internal parameters to minimize the difference between the text it generates and the text it is trained on. This allows the model to learn patterns in the data and generate text that resembles human language.
AI is already replacing humans.
AI has already replaced humans in several fields and will likely replace them in many more in the near future. Here are some examples:
· Manufacturing: AI-powered robots and machines have replaced human workers in many factories, particularly in the automobile and electronics industries.
· Customer service: AI chatbots have replaced human customer service representatives for many companies. They can handle a large volume of simple queries and provide 24/7 assistance.
· Data analysis: AI algorithms can process large amounts of data much faster than humans and provide insights that were previously difficult or impossible to uncover.
· Finance: AI is being used in the financial sector to analyse data, detect fraud, and make predictions about the stock market.
· Agriculture: AI is being used to optimize crop yield and improve farming efficiency. For example, drones equipped with AI algorithms can identify and spray only the weeds, reducing the need for manual labour.
· Healthcare: AI is being used to assist doctors in diagnosing diseases and to help researchers develop new treatments.
· Transportation: Self-driving vehicles powered by AI are being developed and tested by several companies and are expected to replace human drivers in the future.
These are just a few examples of the many fields where AI has replaced or will replace human workers.
The more interesting fields for AI to conquer are the ones requiring “human connection”. E.g., therapy and psychology.
So far, this wasn’t an option due to the limited capabilities of AI. But if AI is trained on millions of sessions of therapy and learns to analyze and respond as accurately as a therapist, could it replace therapists in the near future?
The ethical answer is no. AI shouldn’t replace human connection. The practice of therapy requires human empathy, compassion, and interpersonal skills that AI cannot yet replicate.
But we already know (from countless examples in today’s societies) that humans do not really care about what is healthy as long as it comes with a big paycheck. In fact, social media today is designed in a way that captures the user for as long as possible with no regards to mental health (read Ten arguments for deleting your social media accounts right now by Jaron Lanier).
This might not be exclusive to therapy. With AI becoming more accurate in mimicking human behaviour, the concept of virtual relationship is not as fictional as it used to be. If AI is believable enough, it will create the illusion of a human interaction.
This can contribute to social isolation, as people may be less likely to engage in face-to-face interactions and may rely on AI for their social needs. This can result in a decline in the quality and quantity of human connections, potentially leading to feelings of loneliness and disconnection.
Some studies already suggest that loneliness and social isolation have increased in recent decades, particularly in industrialized countries. This may be due to factors such as the rise of technology, which can provide easy access to information and entertainment but can also foster a sense of disconnection from others. The trend towards individualism and the decline of community-focused activities and institutions may also contribute to loneliness. AI might encourage more and more disconnection.
Humans require connection for a healthy life. Historically, we evolved in packs and communities for thousands of years; breaking this connection might have catastrophic effects. But as said earlier, capitalism doesn’t really care as long as it sells.
Even though this sounds like science fiction (watch Ex Machina – great film), AI is already being used in the dark corners of the internet to create adult visual content and is largely replacing human intimacy. Is it so ridiculous to think it’ll replace friendships or love relationships?
The digital afterlife.
The idea of being able to talk to loved ones who have passed away has always been a fascination for many people. With the advancements in artificial intelligence and natural language processing, this idea might become a reality.
If all goes according to the previous dark predictions, we might aim to take another step against nature and try to conquer death using AI by creating replicas of our deceased loved ones that can digitally live forever.
To create the illusion of talking to a dead loved one, the language model (ChatGPT or a more advanced model that might exist in the future) would be trained on the specific individual's writing and speaking style. This would be done by collecting a large dataset of the person's writing and speech patterns, including the way they use language (specific words and phrases they use in daily life), grammar, and punctuation.
Once the algorithm has been trained on the individual's data, it would then be able to generate text that mimics the person's style of talking. This would allow the user to have a conversation with the individual as if they were still alive.
While this technology may seem like a way to bring closure and comfort to those who have lost a loved one, it is important to keep in mind that it is not a perfect substitute for the real person. The algorithm may be able to generate text that resembles the individual's speaking style, but it does not have the person's thoughts, feelings, or emotions.
The advancements in artificial intelligence and natural language processing might make it possible to talk replace relationships or even our loved ones who have passed away. The technology behind ChatGPT might be the first piece of the puzzle to allows us to train the algorithm on a specific individual's behaviour, which in turn enables us to have a conversation (or a relationship) with a virtual friend (or a deceased relative if they were still alive). It should be noted that there exists already AI models which generate visuals (images and videos) and even audio mimicking a real human being.
However, it is important to remember that this technology is not a perfect substitute for the real person and should be approached with caution.
Thanks to ChatGPT, for helping me write this article (I know, I know... 😊).