Share

ChatGPT, what it is and how artificial intelligence that answers people's questions works

What is the application of the moment and how does it work? The first bans on use. The cyber war between Microsoft and Google

ChatGPT, what it is and how artificial intelligence that answers people's questions works

In the last month we have been talking about Chat GPT, an application that, taking advantage of the increasingly advanced artificial intelligence, answers user questions automatically generating texts.

What is ChatGPT

ChatGPT (acronym for Generative Pre-trained Transformers) it's a'chatbot application launched in November 2022 by OpenAi, a non-profit organization based in San Francisco. The goal of the organization is artificial intelligence (AI) research with the purpose of promote and develop a friendly AI so that humanity can benefit from it.

ChatGPT is avirtual assistant created by GPT-3 language model, a sophisticated model of machine learning of natural language processing it uses advanced algorithms of machine learning to generate responses similar to human natural language.

Thanks to artificial intelligence, therefore, ChatGPT allows you to dialogue (chat) in written form, in a very realistic way, with people receiving an answer in seconds: the application is able to answer questions, create articles, generate texts and so on. He is also able to solve mathematical operations and review programming codes. The application does not answer questions concerning conscience or feelings or those that are inappropriate.

It's the same application, by asking her directly, to explain what is: “ChatGPT is a template of natural language processing based on a neural network that was it trained on a large amount of text available on the internet to generate text autonomously and answer user questions in a natural way. It was developed by OpenAI and has been used to generate text, conversation, translation, and other natural language processing applications. ChatGPT uses a learning transfer technology that allows it to adapt to a variety of domains and to use his training for new tasks.

Un modern virtual "robot". ready to help us in our requests.

How ChatGPT works

ChatGPT is accessible to all and its use is very simple. The first thing to do is go to the site chat.openai.com and sign up for free, creating your own account. The interface is very clear and intuitive. On the homepage there is a introductory guide and examples that can be used for requests.

Once the chat page is launched, you can start interacting with the AI. In the last period, given the enormous use on a global scale, it is sometimes necessary to wait for some time before being able to use it. In the chat page there is a box where you can enter your request; ChatGPT processes the entered input and promptly generates a coherent response to the entered message. Yes you can ask anything to the application: for example if we ask him to “write an article on the Italian economy” he will start creating a text in a few seconds. The work can be generated in any language, simply asking him to write it in the desired language (e.g. write an article in English on FIRSTonline).

Le opportunities for use are as varied as possible: you can use ChatGPT to ask him to write a song or to create an alternative text, generate a rhyming poem, create conversations between people, generate physics formulas, make summaries similar to those of children, translate programming code. Practically, ChatGPT can do virtually anything that can be communicated through writingas far as possible and permitted.

Chat GPT however it is not perfect: sometimes in drafting the text he commits mistakes or blunders. Another thing it can't do is search recent online information because its database is updated until 2021. If you ask it what happened in 2022, the AI ​​will answer like this: “Sorry, I have no information on 2022 as my knowledge has been cut to 2021. I suggest you consult trusted news sources for information on recent events.”

It must be remembered that ChatGPT is not an oracle, it is an artificial intelligence but it was trained by researchers. Behind the development process there are people which they showed toAI how to interact, choose the most effective dialogues and which were the best answers. The idea behind ChatGPT is not to leave the artificial intelligence free to choose the best answers autonomously, but to progress, through people's feedback, to improve its use.

Usage issues

Interest in ChatGPT has begun to raise many doubts about its use. Many have begun to use ChatGPT, mostly young students looking for solutions for school homework. Many discussions have arisen around its use, so much so that some have already intervened forbidding its use.

Il New York Department of Education banned the use of ChatGPT in schools. The schools of the Big Apple, in fact, have blocked theaccess to the platform from their institutions. The outright ban stems from fears that the application could have a negative impact on pupils' learning. In Australia instead, the university have returned to perform the pen and paper exams to prevent the chatbox from helping students with homework. A no shortcuts policy for students. A is being born duel about who supports artificial intelligence as an exaltation of human creativity and whoever believes this is its end. A modern confrontation between values ​​and technology.

Il New York Times exposed himself by defining the application a danger to democracy with fears that artificial intelligence could replace human minds. For Bloomberg, ChatGPT is the start of the artificial intelligence revolution that will profoundly change the economy and the way of life of the people. The moral and ethical discussion on the use of this technology is just beginning.

Meanwhile for counter the use of ChatGPT applications are emerging that recognize whether the text has been written by an automatic system. At the moment there is no official solution to prevent the use of this technology. The next challenge in fact it will be that of create tools that recognize whether or not the generated text was created by an artificial intelligence. The same OpenAI may decide to insert a watermark invisible to generated content. At the moment there are only a few tools to try to understand if the text is generated by AI: just enter a text to analyze it and understand the probability that it was created by an artificial intelligence.

The fight between Microsoft and Google

Microsoft has long been sensing the business of artificial intelligence. The Redmond company already has invested 1 billion dollars in 2019 in OpenAi but now she seems to be intent on a investment of 10 billion dollars. Microsoft's goal is make ChatGPT accessible from your search engine Bing in order to counteract the global domain of Google. If the application were integrated, in fact, this could lead to the end of Google's monopoly. In fact, searches on Chrome could be rendered useless by AI capable of overriding search engines.

Microsoft already had theaccess to technology by OpenAI. A few days ago he decided to open this technology to his parents Azure customers, after allowing access to a limited number of business customers starting November 2021. But that's not all because Microsoft wants integrate ChatGPT into your Office package (Word, Excel and PowerPoint) thus promising a revolution in writing and operating for its clients.

GoogleInstead, after losing the battle to acquire the technology, he promised war on AI-generated content. The Mountain View company has promised to create a tool that can recognize the generated content with artificial intelligence. Google has already created a Chrome extension, called Detect GPT, which indicates whether or not the text of the page we are visiting has been generated by artificial intelligence. The fear of the IT giant is that artificial intelligence could compromise its business; danger that led the company to recall its founders, Larry Page e Sergey Brin, to discuss new tools and solutions.

comments