What is ChatGPT, the artificial intelligence that writes like a human being (and why it’s scary)


What is ChatGPT, the artificial intelligence that writes like a human being (and why it’s scary)
Written by aquitodovale

Open Ai the company created by Elon Musk

It was Elon Musk who created the phenomenon of the web. He was among the first to believe in Open Ai, a non-profit organization that studies artificial intelligence with the aim of promoting and developing friendly AI software so that humanity can benefit from it. It was founded in 2015 in San Francisco and its patents are open to the public. It has several products but the most famous has become the ChatGPT chatbot. However, after the investment announced by Microsoft, the developers of the model are thinking of putting everything up for auction with a monstrous starting point: 29 billion dollars.

What is a chatbot

For some years you have probably already interacted with an artificial intelligence. Large companies already use what are chatbots, i.e. software designed to simulate a conversation with a human being. Currently they are able to answer a few questions and then perhaps refer you to the real human being of the assistance service. ChatGPT is a more powerful version of a chatbot.

What ChatGPT does

And so we come to Chatgpt: what is it able to do? Can create text like a human, using clear, crisp prose and appropriate punctuation. To do this, it uses 170 billion parameters and a database of 570 gigabytes of data. This chatbot not only answers questions but can also write entire texts using different styles: all based on the parameters that the user is able to provide. Questions like “write me a love story between Renzo and Lucia set on the Moon using the style of Shakespeare” could generate masterpieces of fantasy literature. ChatGPT was launched in November (here the link where to try to register) and since then millions of people have signed up. Such a success that, at the time of writing, it is currently impossible to join the community.

Fake news, plagiarism: what are the risks of ChatGPT

But like any medal, there is a flip side. Having an AI that can write like a human could “deplete” humanity. The risk of seeing the web and social networks invaded by fake news that appear to be real (even scientific) texts is very high. And even in schools an invasion of term papers “made” on the computer is expected. Without forgetting the artistic side: many songwriters fear the arrival on the market of products similar to their creations. And how many jobs could therefore disappear (including journalism) with the arrival of an even more powerful Ai?

The race for cover to defend against wrong use

However, what must be remembered is that the software works when at the base there are original data created by man anyway. If everything begins to be produced by the machine, in the end even the underlying data will always remain the same. And what about all those professions that will be created in the future: data scientists, machine learning engineers or Ai developers. And at the same time there are systems on the way that are able to use artificial intelligence to hunt down written texts… with artificial intelligence. ChatGpt Zero is an app to combat plagiarism in schools and universities, which have increased since ChatGpt was released, and was created by Edward Tian, ​​a 22-year-old computer science student at Princeton. It is a kind of “antidote” to the AI ​​system that has become popular in recent weeks which is however generating doubts in the academic world, IT security and also in the artistic world for issues of creativity and copyright. The GptZero system was created by Edward Tian to “quickly and efficiently decipher whether a human or ChatGpt wrote an essay”. The app analyzes certain language parameters to check if a text was written by a person or by the AI. For example, it measures its complexity (the higher it is, the more likely it belongs to a person) or the variation of periods (an artificial intelligence tends to write with a uniform style).

The risks of ChatGPT in schools

The system developed by the student is grafted into a debate on the texts and conversations produced with the ChatGpt software that various schools and universities, from New York to Australia, have begun to prohibit because their construction and syntax are difficult to distinguish from human ones and therefore at risk of student fraud. A policy also shared by the International conference on machine learning which has prohibited the use of these systems for writing academic papers. Also alarmed by cybersecurity experts because ChatGpt is capable of writing malicious viruses. Concerns relating to the ethics and correctness of using this software also affect the field of art for issues of plagiarism and copyright infringement. “Crap” and “a grotesque mockery of what it means to be human”. This is the opinion of Nick Cave, musician and composer, on a song written by ChatGpt “in his style” and submitted to him by a user. An exercise in “replication and parody,” he added, “writing a good song isn’t imitation or replication, it’s the other way around.”

The future? An even more powerful ChatGPT

However, the technique does not stop and indeed an even more powerful version 5 of ChatGPT capable of using trillions of parameters and terabytes of data is on its way, even if the founders continue to deny it.

And the competition doesn’t stand idly by: that’s what Google does

Certainly Artificial Intelligence is a tempting market and cannot be left in the hands of a single company. And it is for this reason that Alphabet (Google) also moves its pawns. On the other hand, their main product, the search engine, is in “danger”. This was revealed to Time by Demis Hassabis, head of DeepMind, the AI ​​company headed by Alphabet. He explained that the company is developing a trial, private version of its Sparrow chatbot that could be released in 2023. Hassabis says this software has features that ChatGPT lacks, including the ability to cite sources through reinforcement learning. DeepMind and OpenAI are both working on the achievement of artificial general intelligence (Agi), which has the goal of creating hardware or software that can emulate human reasoning as closely as possible. “The creation of Agi will be the most important technological development in human history,” said OpenAI CEO Sam Altman.

#ChatGPT #artificial #intelligence #writes #human #scary


About the author


Leave a Comment