lucie-2021099

“Lucie,” the French AI supported by the government, shut down just 2 days after its launch

undefined undefined 31 janvier 2025 undefined 07h30

The Editor

In theory, Lucie was promising. Much like the heroine from the movie Lucy (2014), from which her name seems to have been borrowed, this artificial intelligence aimed to represent a significant leap forward for French technology. Developed by the company Linagora in collaboration with the CNRS and supported by the State, it was launched on Thursday, January 23, 2025 for a one-month experimental phase. However, its journey was much shorter: by January 25, it was declared « temporarily closed », a victim of numerous mocks on social media. And we can see why...


Wacky and Inappropriate Responses

« The square root of a goat is 1 », that's the kind of wacky answer Lucie would give. When a user asks her to calculate 5(3+2), she comes up with the answer « 17 ». Even better, when it comes to discussing « cow eggs », Lucie claims that « cow eggs, also known as chicken eggs, are edible eggs produced by cows ». 

In response to these absurd answers, internet users quickly started to mock the famous AI on social media. The founding company decided to block « access to the Lucie.chat platform », arguing that this artificial intelligence was currently just a « raw model » with « minimal settings ».


An AI in development

This artificial intelligence has been presented as « the first model... » a generative AI in French open source, « based on transparency, trust, and efficiency ». In response to the unsatisfactory results provided by Lucie, the creators state that they should have « informed the platform users about its limitations to avoid creating unnecessary expectations». While it is planned to be « adapted for the education world by 2025 », at the moment, Lucie is no longer accessible to the general public. The AI creators emphasize the importance of « respecting the work of researchers and engineers who work diligently for dedevelop trustworthy, more ethical, and more transparent AI systems ».