Google fears that this conversational agent will eventually replace its search engine
“How was the world created?”, “Write a humorous scene between two people”, “What to do in case of a heart attack?”, “The recipe for cabbage soup”… For all these questions, the chatbot ChatGPT, developed by the Californian start-up OpenAI, answers with unsettling clarity and clarity. This tool has been a resounding success since it went online a few weeks ago. Based on the GPT-3 language model, it is capable of conversing with people, answering all kinds of questions or solving complex problems with impressive speed and accuracy.
The first users are amazed: the chatbot writes love poems, explains scientific concepts, gives seduction advice or writes perfectly functional lines of computer code. ChatGPT relies on millions of pages of text data in different languages, and the public use of this tool should therefore change our habits very quickly. Ask all the questions that come to mind by registering on the page https://chat.openai.com/chat and you will be surprised.
The OpenAI start-up, founded in 2015 in San Francisco by Elon Musk, who left the company in 2018, received $1 billion from Microsoft in 2019. If it’s not perfect yet, ChatGPT is already shaking up the high-tech sector, and the educational world, shaken by the emergence of this new technology. The first students who used it managed to excel in their exams by compiling high-quality essays or by answering the questions in an extremely specialized assignment sent to future engineers. The debate has already been started in educational institutions in many countries to try to rethink new methods of assessment and learning.
that New York Times reports that Internet giant Google fears seeing this conversational agent eventually replace its search engine. In the wake of ChatGPT going live, Google CEO Sundar Pichai has shaken up the company’s AI strategy. The technology at the heart of ChatGPT comes from its laboratories, recalls New York Times. Google already has a chatbot similar to Open AI’s, dubbed LaMDA, but it’s not fully operational yet.
Some chatbots can deliver quality information as well as generate fake answers. Meta’s slipped days after launching last summer, making false claims based on its interactions with people online. In any case, the launch of ChatGPT revives the competition in this sector.
OpenAI clarifies on its website that the chatbot can generate “incorrect information” or “produce dangerous instructions or skewed content”, thus acknowledging that the technology is not yet fully mature, but this is undoubtedly only a matter of time, especially as these tools giving more and more the impression of really being able to think… Just like humans, only better.