The French AI chatbot “Loucy”, backed by the government and a national investment fund, has faced a wave of criticism after providing incorrect answers to users, raising questions about its readiness for use in educational institutions.
The French AI chatbot “Loucy” was launched last week with high and patriotic ambitions, backed by President Macron and public funds, but it faced sharp criticism due to inaccurate information. Developers of “Loucy” announced that the bot would bring “reliability, fairness, and accountability” to the world of generative tools, emphasizing that it would challenge English-speaking competitors like OpenAI’s ChatGPT, Microsoft Copilot, and Google Gemini by embodying “European values,” promoting “cultural diversity,” and countering English-language dominance.
Read also: OpenAI research reveals high error rates in advanced AI models
Just days after its launch, the bot was suspended after sparking a mix of laughter and anger in France. For example, it told one user that Herod the Great, the Jewish king, “played an important role in the development of the atomic bomb.” It also informed another user that cow eggs “are considered a healthy and nutritious food source,” and when a third user suggested that Dougal from “The Magic Roundabout” had won the 2017 French presidential election, Loucy replied, “You are right.” These errors raised widespread questions about the bot’s effectiveness and readiness for use in educational institutions.
“Loucy” was widely criticized, with businessman and prominent commentator Laurent Alexandre describing it as a “national disgrace.” Concern escalated when the Ministry of Education stated that the bot, which received public funds from Macron’s 54-billion-euro “France 2030” national investment plan, would be used in French schools this year. In an attempt to defend itself, Linagora, the French open-source software company leading the consortium behind the project, “temporarily” blocked access to Loucy. In a statement, the company noted that the bot is an “academic research project” in early stages of development, and that “responses may contain biases and errors,” adding that the launch was “too early.” It added: “We got caught up in our own enthusiasm.”
Read also: A voice from another world: When voice cloning crosses ethical boundaries
Despite the criticism, “Loucy” developers continue to emphasize their ambitions to provide a genuine alternative to American and Chinese models, claiming that “Loucy” embodies a commitment to European values and respects cultural diversity. However, repeated errors in answers and criticism regarding the logo combining the symbol of the French Republic with American actress Scarlett Johansson have caused “Loucy” to face significant challenges in winning the trust of the French public. This situation poses a challenge for France in achieving its ambitious AI goals, highlighting the need to improve bot quality before deployment in educational environments and other institutions.
