What is Chatgpt?
With 1 million users in less than a week after its launch, Chatgpt has something to impress. A chatbot catotype capable of talking naturally with human language In one way that no other artificial intelligence could achieve. In addition, it can generate textual conversations with almost human precision. Fyi, a chatbot is a program capable of talking with a surfer and GPT is the acronym for generative transformation prior to training.
The new Chatbot was developed and launched on November 30, 2022 by Operai, a research laboratory in San Francisco. The objective is to promote and develop “An artificial intelligence with a human face that will benefit all humanity”. Operai is an artificial intelligence research organization (AI) whose objective is to generate advanced advances in AI, particularly with respect to ethical and social aspects.
How does the chatgpt work?
The beginning? Ask your questions to ChatgPT and artificial intelligence will instantly generate an answer. The speed at which the tool is put to produce specific responses is quite disconcerting. But how does it work? To answer you The tool will draw your information from a huge 2021 Internet database and many books.
This conversational robot, always in beta phase, synthesizes this information to offer a clear and precise response. If the answer does not fully satisfy you You can repeat your application and the program will answer differently. Designers work continuously to develop the humanization of Chatgpt responses, to answer their natural language questions. To summarize, the program feeds on existing information, but also the comments of the users who can offer an answer themselves.
You can use Chatgpt as help to create websites, develop a mobile application or even mobile video games. Chatgpt can also tell a story, generate text, information and format it. This artificial intelligence deals with all issues: from the search for information to the help of translation through the explanation of historical or scientific facts … however, it is specified that The knowledge of the tool is limited to 2021 And, of course, you will not answer questions about inappropriate requests and illegal issues.
The new conversational uses techniques of Automatic learning o Automatic learning. In other words, this is the ability of machines to learn a set of automatically rules of the data. ChatGPT is animated by the Operai GPT-3 algorithm that uses deep learning or deep learning, based on the neuronal networks model. It is able to write text content considering all codes of natural language. (Grammar, spelling, syntax, etc.). A real help to write and translation in all languages!
The actions of this intelligent assistant are really impressive. A real -time cat application capable of following a conversation is A real revolution.
“Chatgpt is optimized for dialogue. Our goal is to make the systems of AI more natural to interact with him, and his comments will help us improve our systems and make them safer.”OpenAI says.
At the free moment, the chatbot has a very easy to use interface, accessible to everyone around the world and in all languages.
Chatgpt limits
Despite a precision that seems disconcerting, ChatGPT cannot guarantee any insurance for the authenticity of information that he provides. In the era of false news and misinformation, AI without discernment would do it without discernment of false, harmful or inappropriate information and spread them through their conversations. It cannot distinguish the truths from lies. Even creators specify: Chatgpt can provide deceptive information. Although innovative, the solution has defects in the accuracy of certain data.
As mentioned above, The knowledge base ends in June 2021AND Chatgpt does not have Internet access or news. The answers provided are made of large databases of texts recovered from the Internet and, therefore, learn the worst as the best of human expression.
Journalists fear for their work, but especially for the reality of the information communicated to the population. In fact, the chatgpt system could present a certain number of ethical and social risks, such as perpetuate misinformation and raise confidentiality problems. Certain applications are prohibited and Openai’s mission is to moderate inappropriate or even racist and sexist applications; All questions “Discriminatory, offensive or inappropriate”.
OpenAi provides “Our goal is to obtain external comments to improve our systems and make them safer. Although we have implemented protection measures, the system can sometimes provide inaccurate or misleading information and produce offensive or biased content. It is not intended to give advice.”
The plagiarism problem
To treat innumerable sources used by chatgpt (and dall-e for images), we must take into account Copyright problems. The program must also analyze and integrate thousands of online resources. Chatgpt will find problems creating answers because you can plague data. If a student had to use his results in a duty, he could face accusations of plagiarism or find his work cited without authorization in other places. To avoid deceptive readers, The texts generated by the chatgpt will carry a brand that stipulating that they were created by AI and not a humanin particular university plagiarism.
Bing emergency buoy?
The chatbot could be integrated into the bing search engineto weaken your great Google competitor. Microsoft has already integrated another famous AI tool created by OpenAi in its search engine with Image Creator. These are Dall-E, a program capable of creating images based on textual descriptions.
Microsoft expects to give the Bing search engine the opportunity to attract more users and significantly improve their user experience by providing natural and detailed written responses.
Chatgpt in summary
Despite its weaknesses, Chatgpt is a true revolution born from the web and will extend to many uses …
Chatgpt capabilities
- Remember what the user said previously in the conversation
- Allows the user to provide monitoring corrections
- Trained to reject inappropriate requests
- You can create simple programs in all programming languages
Chatgpt limits
- Sometimes you can generate erroneous information
- Occasionally he can produce harmful instructions or biased content
- Limited knowledge of the world and events after 2021
Chatgpt remains The most sophisticated artificial intelligence program And also the most accessible. “My goal is to generate text at the request using maximum artificial intelligence. I am a tool to generate text, not to speak.”said. However, it is important to keep in mind that your answers are not always exact. The current version seems to be a beta version, which must still be improved.
According to Openai, the objective is that general artificial intelligence benefits all humanity, which has Internet access in any case. Ultimately software such as Chatgpt could even replace web services such as Google and Wikipedia. AI is not surprised. Our web agency in Saint-Malo supports it in its digital transition and in its focus with the tools provided by AI.
