ChatGPT Welcome technological advances but understand the limitations of this

ChatGPT: “Welcome technological advances, but understand the limitations of this type of system”

The general public has been fascinated by the power of ChatGPT, this conversational agent developed by OpenAI and recently available on the web for free.

ChatGPT is an artificial intelligence (AI) tool for testing the performance of the latest voice generation models, called Transformers. ChatGPT can be used for many tasks: summarizing a text, answering questions, creating or correcting a computer program, completing a task, or even simulating an interaction with a deceased person.

While everyone plays with and proudly talks about “amazing” experiences, few delve into the ethical aspects of these black boxes. Of course we have to applaud technological advances and remarkable performance, but we also have to understand the possibilities and limitations of this type of system.

Also read: Article reserved for our ChatGPT subscribers, the software capable of writing confusing small texts

What is ChatGPT up to? Texts he digested using algorithms… GPT-3 (Generative Pre-trained Transformer 3) is the language model of ChatGPT, a neural network architecture with 175 billion parameters that uses pre-trained linguistic representations from huge databases (Wikipedia, websites) to produce texts that mimic human speech in multiple languages. Due to its design, the model only recorded information about the words and word components and their context of appearance.

Flooding the world with fake news

ChatGPT is therefore trained to give the following sequence of the most probable words from a text input – which is not really an answer to a question! The model is obviously flexible enough not to always produce the same text. ChatGPT also uses reinforcement learning by human trainers who have evaluated multiple alternative answers. The model is thus optimized by human intervention.

ChatGPT is a socio-technical system that weighs on a variety of ethical values. He can imagine stories that are harmless or caustic: we’ve seen stories of rabbits… or the design of weapons.

The most immediate danger of these systems is the possibility of flooding the world with billions of “fake news” as ChatGPT is capable of text inventions of all kinds when played with. It could then be fed with this new artificial data collected from the internet.

Also read: Article reserved for our subscribers Beyond artificial intelligence, the ChatGPT chatbot owes its language skills to humans

We would end up with incredible processing of opinions, knowledge, bibliographies that might be machine wrong and fuel disinformation on the internet. It’s still easy to get ChatGPT wrong on very specific questions where it’s unable to argue and separate true from false. But you have to learn to question the answers of the system. Banning ChatGPT in school may not be the best solution. learning to use it is certainly a key to accelerating our understanding of these systems and preserving our free will.

You still have 50.33% of this article to read. The following is for subscribers only.