We tested ChatGPT, the computer robot that writes texts… Bluffing and a little disturbing! – Evening edition Ouest-France – 06/12/2022

By Clementine MALIGORNE

Developed by the Californian start-up OpenAI, the ChatGPT software is designed to generate texts. Particularly powerful, managed by artificial intelligence, this computer robot is capable of producing surprisingly coherent and successful writing.

Cooking recipe, poem, scientific concept, political speech, dissertation, children’s story, press article… The performance of ChatGPT, a text generator based on GPT-3 technology, managed by artificial intelligence, is as much fun as they fascinate. This online platform answers your questions and executes your instructions by writing content. Its latest version, launched at the end of November, is particularly successful.

After the success of DALL-E 2, an online software that generates a computer-generated image as realistic as a real photo, based on a few descriptive words, again thanks to artificial intelligence, the Californian start-up OpenAI, co-founded in 2015 by Tesla boss Elon Musk (who has since left the company), launched ChatGPT, a conversational robot at the end of November. (chatbot)able to answer questions and lead a real exchange with a human interlocutor.

Read also: This software creates with one click a hyper realistic image from a simple description

This version is currently freely available on the internet. All you have to do is fill in your email address to access it. On Twitter, Internet users have fun testing the capabilities of this software, which has what it takes to compete with the Google search engine. The requests are each crazier and more unusual than the next: “Write Emmanuel Macron’s New Year’s speech by placing the words ‘kettle’, ‘Jerusalem’ and ‘duck'”, asks a user. Or : “Write a story with Babar, Mata-Hari and Pierre Bourdieu. »

We tested this robot by first asking it a question that could lead to an article. “Why are pharmacies running out of Doliprane for children? » Here is his response: “I’m not sure exactly why pharmacies might run out of Doliprane for children right now. This could be due to a high demand for this medicine due to an increase in cases of illnesses in children, or problems with the production or distribution of the medicine. I recommend that you contact your local pharmacy to find out more about the availability of Doliprane for children. » Thanks for the advice !

The GPT software gives advice and explanations to concrete questions such as: “Why do pharmacies run out of Doliprane?” (Screenshot: ChatGPT)

We also asked him to give us a pancake recipe (which we didn’t try).

ChatGPT software crepes recipe. (Screenshot: ChatGPT)

The robot also skilfully lent itself to a philosophical dissertation on passing time, with introduction, thesis, antithesis and synthesis.

ChatGPT is capable of producing a short three-part dissertation on the passage of time. (Screenshot: ChatGPT)

How to declare your flame? Here is GPT’s answer. “The important thing is to find a way to communicate your feelings sincerely and authentically. » We keep !

The software responds to all kinds of queries. (Screenshot: ChatGPT)

Disturbing performance

Surprising, these performances are also a little worrying. Something to relaunch the debate on the risks associated with artificial intelligence (AI) technologies. Indeed, many observers fear that this type of technology will be misused, by spreading false information for example, or by creating increasingly credible scams.

Others fear that certain professions, such as journalists, teachers, playwrights, advertisers or even computer programmers, whose function is to produce content, will one day be replaced by this kind of intelligent robot.

The British daily The Guardian also reports that academics, who tested the tool by generating answers to exam questions, judged that this GPT-3 would obtain scores comparable to those of a license level… Programmers also used the tool to solve coding problems and the robot delivered an answer to them in seconds. Asked, the ChatGPT-3 recognizes it itself: “It is true that computer programs like ChatGPT-3 may cause concern because of their ability to simulate human conversations very realistically. »

However, safeguards have been put in place to prevent abuse. ChatGPT, for example, refuses to take sides on opinions or controversial subjects. “It is important to keep in mind that these types of programs are not perfect and cannot have independent thoughts and opinions, and are not able to completely replace human interactions,” he answers us.

This tool, while powerful, is not infallible. On the home page, OpenAI also specifies that the chatbot can generate “incorrect information” Where “produce dangerous instructions or biased content”. And to conclude: These computer programs can be useful for certain tasks, but they will never completely replace human relations. »

A version 4 would be in the pipes, but it has not yet been officially announced.

Chat GPT-3 acknowledges that such programs may cause concern as they simulate very realistic human conversations. (Screenshot: ChatGPT)

We would love to say thanks to the writer of this short article for this amazing material

We tested ChatGPT, the computer robot that writes texts… Bluffing and a little disturbing! – Evening edition Ouest-France – 06/12/2022

Take a look at our social media accounts and also other related pageshttps://yaroos.com/related-pages/