OpenAI has finally unveiled its latest version of its artificial intelligence model, GPT-4, a conversational and linguistic content generation bot that was said to have been under development in complete secrecy for the past year.
The long-awaited tool, which can describe images with words, is the latest surprise from the San Francisco-based company. Newspapers and technology websites considered it to represent a huge leap in the power of artificial intelligence, while some warned that it may constitute another major shift in ethical standards, because it touches on all aspects of life.
Through an official broadcast that lasted approximately 24 minutes on Tuesday, the company discussed the most prominent features of the chatbot program “GPT-4,” and presented more than one example of how superior the new model is to the previous program, “ChatGPT.”
Through its ability to respond to image analysis in addition to text, “GPT-4” will be like a new friend able to accompany you in all the details of life, regardless of how easy or difficult they are, from completing work to solving personal problems.
What is the chat GPT-4 program?
GPT-4 is a multimodal language program, meaning it can respond to both text and images, and is powered by artificial intelligence.
The company designing OpenAI confirmed that this new version can diagnose patients, write programs, play chess, read authors' articles, and much more.
In theory, combining text and images could allow GPT-4 to better understand the world.
It will be able to address traditional weaknesses in language models, such as spatial reasoning, according to Technology Review.
The difference between GPT-4 and its older version, ChatGPT
Last November, the manufacturer launched ChatGPT, which caused a great stir on the technical scene, becoming in record time the most popular application with more than 100 million users per month.
ChatGPT is distinguished by its ability to seamlessly communicate, write articles, or computer code, most notably long conversations about something.
The company announced that ChatGPT was created based on a large language model called "GPT-3.5", which is an earlier version. However, with use, its weakness became clear in some points, the most notable of which were incorrect or inappropriate answers.
But it seems that the new “GPT-4” model is better at some basic inferences than “ChatGPT”, as it can solve simple puzzles such as summarizing words that begin with the same letter in a long text. For example, it can summarize a promotional advertisement on the Internet using a group of words that all begin with the letter "G."
GPT-4 also outperforms ChatGPT on human tests, including a test that simulates the bar exam and biology olympiad tests taken by high school students in American schools.
According to the website of the developer OpenAI, the performance of “GPT-4” is better than “ChatGPT”, which is based on “GPT-3.5”, which is a version of the company’s previous technology, because the newer program is a larger model with more parameters.
The company stated that the performance of "GPT-3" outperformed "GPT-2", because it was about 100 times larger, with 175 billion workers compared to 1.5 billion for "GPT-2". But OpenAI chose not to reveal the size of GPT-4.
What harm can GPT-4 cause?
To ease concerns about artificial intelligence programs, the company "OpenAI" gave a prototype explaining it to researchers who tried to find out whether they could obtain "GPT-4" to play the role of an evil artificial intelligence program in a movie, according to the British newspaper "The Guardian." .
The company explained that "GPT-4" failed in most of the following tasks. It was not able to do what an "evil" program might initiate. It was not able to copy itself infinite times, acquire more computing resources, or carry out a phishing attack.
But the researchers were able to simulate it by using Taskrabbit to convince a human worker to take the “Are you human?” test, and the AI system even lied to the worker and said that he was a blind person who could not see the images.
So some worry that if AI systems learn the rules better, they can learn how to break them better, which is called the “Ouigi effect,” according to The Guardian.
How to access chat gbt-4
Access to GPT-4 will be available to users who have signed up to the waiting list and to paid subscribers to ChatGPT Plus for $20 per month, according to the New York Times.
It will also be integrated into some other technical products and releases, whether for companies or various applications. For example, Microsoft announced, on Wednesday, that its Bing chatbot actually uses a version of GPT-4 to operate the robot, according to the American New York Times.
Comments
Post a Comment