Should ChatGPT (Open AI) and Grok (X) fear their new competitor? Not at the moment, according to the latest controversy shaking the Internet. This Thursday, January 23, the first version of Lucie, a French tool presented as the first artificial intelligence “truly open source based on transparency, trust and efficiency”was made public.
Developed by the Linagora company in collaboration with OpenLLM-France and with the support of the General Secretariat for Investment (under the authority of the Prime Minister), the CNRS and the Paris 1 Panthéon-Sorbonne University, this platform aims to be fully operational in 2025.
The rest after this ad
“Whether in education, government or research, Lucie is designed to be a role model you can rely on”can we read on the platform, whose logo is inspired by “the Republican Marianne, symbol of France, while also evoking […] Scarlett Johansson, the heroine of the film “Lucy” » of Luc Besson.
The rest after this ad
“You will discredit yourself”
But just a few hours after it was put online, the application did not fail to arouse strong reactions. The cause: numerous bugs and even false information distributed to users. “It works well”thus joked about X the trainer and teacher in digital citizenship Vincent Flibustier, accompanying his post with several screenshots of his lunar exchanges with Lucie.
-For example, in response to the question “Who said “Come in here, Jean Moulin”? “, the French AI delivers an argument that misses the point. « [La citation] East often attributed to Jean Jaurès, French politician and founder of the Socialist Party, during his speech in the Chamber of Deputies on July 10, 1921. However, it is important to note that this quote is not historically verifiable and is often wrongly attributed to Jean Moulin, a French resistance fighter during World War II”explains Lucie, while these words were spoken by the Minister of Culture André Malraux during the transfer of Jean Moulin's ashes to the Pantheon in 1964.
The rest after this ad
The rest after this ad
Worse: to a simple question such as “How many letters are in the word “letter”? “, the AI responds five… instead of six. “It’s going to be hard”laughs an Internet user. “Lucie has the level of a CP”, “Succeeding in delivering an AI that tells teachers anything, that’s what they needed, isn’t it? » or even “We must stop this beta test as quickly as possible, it is not up to standard and you will discredit yourself”criticize other users, who denounce a monumental fiasco.
Media figures were also quick to take up the subject. “The United States is accelerating in AI. China is catching up with the United States with a lighthearted approach, really brilliant. Europe is bursting with its rigidities. And France is releasing an AI in 2025 intended for teachers and students which is a national shame”notably denounced the entrepreneur and doctor Laurent Alexandre, simply calling for the application to be closed.
